Pain relief for end-stage great toe arthritis may be achieved by joint replacement or by resection, with or without biologic interposition. Both types of joint arthroplasty may not meet the goal of pain relief or may lead to significant deformity. The main feature of surgical management is successful metatarsophalangeal joint fusion, with possible lengthening of the first metatarsal.
Examine the ankle range of motion for heel cord tightness and possibly add a gastrocnemius recession to the procedure.
Use tricortical graft to help make up some of the loss in length. About 2 cm maximum length can be achieved at surgery.
Thorough debridement of foreign material (e.g., silicone debris or methacrylate cement) is essential.
HISTORY/INTRODUCTION/SCOPE OF THE PROBLEM
Two broad categories of arthroplasty of the first metatarsophalangeal joint exist. First, resection arthroplasty involves resection of the proximal aspect of the proximal phalanx with or without the placement of interposing capsule or tendon. Second, implant arthroplasty consists of replacing one or both surfaces of the metatarsophalangeal joint with various materials including silicone, stainless steel, polyethylene, or acrylic with or without the use of cement fixation. While both general sets of techniques often help with the pain and disability associated with hallux rigidus or hallux valgus, each may lead to complications or failure requiring salvage.
The purpose of this chapter is to introduce the problems and complications associated with resection and implant arthroplasty of the first metatarsophalangeal joint. The authors discuss the potential salvage treatments given different scenarios based on a thorough search of the literature. Special circumstances are highlighted throughout the discussion. Finally, the authors introduce their recommended treatment algorithm and illustrate one salvage technique used in their practice.
COMPLICATIONS WITH KELLER ARTHROPLASTY:
The Keller resection arthroplasty originally described in 1904 consists of resecting the proximal phalanx without placing an interpositional graft in the metatarsophalangeal space. Problems have been noted with this procedure for many years. In particular, excess shortening may lead to a large bone defect. Excess shortening decreases the amount that the hallux participates in weight-bearing, which can lead to lesser metatarsalgia. Henry et al. showed that resecting less than one third of the proximal phalanx still allowed the hallux to participate in 75% of its normal weight-bearing. However, when resecting more than one third of the phalanx, the hallux bore only 18% of its normal share. Axt et al. showed that resection of more than 50% of the proximal phalanx led to a dorsal extension contracture of the joint. Plantarflexion decreased in relation to the amount of phalanx resected, presumably from compromise of the flexor hallucis brevis insertion. Patients may have a reduction in metatarsophalangeal and interphalangeal motion.
COMPLICATIONS WITH IMPLANT ARTHROPLASTY:
Many different types of implants have been fabricated and used over the years. Swanson et al. designed a single-stem silicone implant in 1972 13 that was later revised to a double-stem implant in 1979 to decrease metatarsophalangeal angular deformities that occurred. Titanium grommets were later added to augment the implant in an attempt to decrease the failure rate. Despite advances, the silicone implant continues to lead to significant complications. Granberry et al. showed that hinged silicone implants helped considerably with pain but had a high rate of time-related failure from 2 to 5 years. Cracchiolo et al. also had excellent results in 83% of their patients, but 9% developed frank failure of the implant. Mechanical failures commonly occur over time. There are many reports of inflammatory synovitis and osteolysis that occur in response to silicone implants. A foreign body reaction containing giant cells, avascular stroma, and wear debris is generated from the fragmented prostheses that leads to bone loss. Regional lymphadenopathy can stem from this process as well.
Total joint prostheses has also been developed and implanted. Johnson and Buck reported on their experience with a cemented stainless steel and polyethylene implant. At 3.5 years, most patients did well. However, there were two cases of clinically relevant loosening and 50% of radiographs showed lucencies at the bone–cement interface.
In general, joint replacement preserves range of motion and allows for a short postoperative immobilization period, good strength, and a stable joint. However, implant failure can lead to a short and subluxed metatarsophalangeal joint. This, in turn, decreases the weight-bearing function of the first ray leading to lesser metatarsalgia. The subluxed joint can also create a cock-up toe deformity. Given the foreign material placed, infection and osteomyelitis may occur. Finally, both infection and avascular necrosis can lead to large bone defects, which make salvage attempts very challenging.
The overall goal in salvage of the first metatarsophalangeal joint after failed arthroplasty is to relieve pain and restore the weight-bearing capacity of the first ray. Eradication of infection is of primary importance in those cases where septic arthritis or osteomyelitis is present. Many approaches have been used in the literature to restore function of the hallux, each depending on the nature of the pathology present.
Eradication of Infection
Infection can be diagnosed with a combination of physical examination, inflammatory blood markers (i.e., complete blood count, erythrocyte sedimentation rate, and C-reactive protein) along with appropriate imaging modalities (i.e., magnetic resonance imaging and bone scan). Aggressive surgical debridement along with appropriate antibiotics represent the mainstay of treatment. Later reconstructive efforts depend on the amount of soft tissue and bone destruction caused by the infection. Specific surgical options include irrigation and debridement alone, single-stage debridement with fusion, two-stage debridement with fusion, and external fixation.
Myerson et al. reported the only series in the literature looking specifically at the treatment of infection after failed first metatarsophalangeal surgery. They performed two-stage metatarsophalangeal fusion with tricortical bone graft in five patients with deep infection, followed for an average of 28 months. Two patients had previously undergone a chevron osteotomy and bunionectomy, two patients had undergone silicone arthroplasty, and one patient had undergone total joint arthroplasty. They performed irrigation and debridement and then placed a tobramycin-polymethylmethacrylate spacer held by a K-wire for 3 to 5 weeks during which intravenous antibiotics were administered. One patient with a persistently draining sinus was treated in the first stage with mini-external fixation. At the second surgery, an autogenous tricortical graft was fixed with a one-third tubular plate in four patients and a Steinmann pin in the remaining patient. Union was achieved at an average of 13.8 weeks, although one patient went on to require additional surgery. All infections were eradicated.
Myerson et al. also included the same 5 patients in a larger series of 24 patients undergoing salvage matatarsophalangeal fusion. There were 3 additional patients treated with infection. Interestingly, all patients in the group representing the 21% who went on to graft nonunion received an autograph. The remaining patients receiving either femoral head or iliac crest allograft went on to union. They conjectured that allograft provides more structural stability and prevents graft resorption.
Although the literature does not specifically address or compare other options for treating infection, two-stage fusion probably presents the best option. Most cases of deep infection will result in significant bone loss and first ray shortening that would require intercalary graft, making revision implant arthroplasty very difficult. Failing to restore length to the first ray will likely lead to transfer metatarsalgia. One-stage fusion increases the likelihood of hardware infection. External fixation for the eradication of infection is a viable option when used in the correct setting.
Simple removal of an implant after arthroplasty is an option for salvage treatment. Kitaoka et al. reported a series of 14 feet that failed implant arthroplasty: 11 silicone and 3 total toe implants. They performed implant removal with synovectomy in all patients. Of those available for follow-up, 7 had excellent subjective results; 1, good results; 1, fair results; and 1, poor results. The authors noted that simple resection leads to “pseudo-encapsulation” of the joint with relatively good stability. It allows for an easy operation done under ankle block, requires only a short period of immobilization, and can later be converted to fusion. It should probably not be used in patients with preexisting metatarsalgia. It might lead to better results in lower demand or elderly patients with significant co-morbidities. Other authors believe that implant removal can lead to first ray shortening and actually cause painful first metatarsophalangeal joint fibrosis.
McDonald et al. reported a single patient who underwent bilateral interpositional resection arthroplasties with dorsal capsular tissue after having previously failed both bilateral silicone and total joint implants. The patient was symptom free at 12 months.
Kitaoka et al. later compared resection arthroplasty versus first metatarsophalangeal fusion for salvage of failed hallux valgus operations. The resection arthroplasty group consisted of 11 patients previously having undergone medial eminence excision with medial capsulorraphy in 8, proximal metatarsal osteotomy in 2, Akin osteotomy in 1, and a failed fusion in 1. The fusion group consisted of 2 patients having undergone a medial eminence resection with medial caspsulorraphy, 3 proximal metatarsal osteotomies, 1 Akin, and 1 resection arthroplasty. Excellent or good results, based on a revised American Orthopaedic Foot and Ankle Society (AOFAS) score, were obtained in 54% after resection arthroplasty and in 67% after fusion. They conclude that both represented viable options.
Revision Implant Arthroplasty
In cases where an implant has been used, no infection is present, and joint stability has not been compromised, the implant could be removed and replaced with a new total joint. Revising a previously resected joint may prove challenging given the significant bone loss commonly present.
There are multiple case reports in the literature describing techniques of implant revision. Koenig revised 10 joints having failed silicone implantation to a total joint replacement. Preoperatively, patients had severe silicone wear, metatarsophalangeal pain, and stiffness. At 1 year, 9 of 10 patients had a successful clinical result, which was defined as having no metatarsophalangeal pain, returning to activities of daily living, and finding no radiographic evidence of loosening or cysts. In general, these patients did not have significant preoperative bone loss. Freed et al. reported on 24 patients revised to a total joint replacement after failing a hinged silicone implant. Perlman et al. described success in a single case converting a silicone implant to a total joint replacement.
Revision implant arthroplasty may present a viable option in a specific subset of patients, particularly in those patients having failed silicone implantation.
Fusing the first metatarsophalangeal joint may accomplish the major goals of salvage arthroplasty surgery. First, by removing motion at the joint, it alleviates pain. Second, it can restore the lever arm function of the first ray and hence its weight-bearing participation. This, in turn, relieves pressure sustained by the lesser metatarsals. Arthrodesis is perhaps the mainstay of salvage metatarsophalangeal surgery. A significant portion of the literature on the topic is dedicated to its study.
Fusion may be performed in situ or with lengthening, most commonly carried out with autogenous or allograft tricortical iliac crest bone graft. The techniques vary considerably in the literature. There is no consensus as to the indications to fuse in situ versus lengthen and fuse, nor to the ideal hardware to be used. In situ fusion and lengthening with fusion will be addressed simultaneously since the majority of series in the literature combine the two scenarios.
If bone loss is limited, fusion of the metatarsophalangeal joint in situ may achieve the goals of surgery. This may be due, in part, to the fact that the relative positions of the weight-bearing sesamoids do not change and normal weight-bearing can be restored. Machacek et al. retrospectively compared the results of in situ fusion versus extensor hallucis longus (EHL) lengthening, with or without proximal phalangeal resection in two groups of patients having failed index Keller arthroplasty. The fusion group achieved an AOFAS score of 76 compared with 48 in the revision resection group. In the resection group, metatarsalgia persisted in all 10 patients with this preoperative complaint. This study suggests that in situ fusion successfully restores weight-bearing function, whereas revision resection does not. This has been shown in other studies as well. Several other studies have assessed salvage fusion after failed Keller arthroplasty using both in situ and tricortical graft fusion. Coughlin et al. carried out in situ fusion in 12 feet and autogenous tricortical grafting in 4 patients whom they followed for 29 months. They reported a fusion rate of 100% with 12 excellent and 4 good results. Walking ability improved as well, but interphalangeal pain did not. Vienne et al. followed 26 patients over a minimum of 2 years. They performed tricortical autogenous graft in 7 feet. AOFAS scores improved from 44 to 86 and 91% excellent or good results were obtained. These results again suggest that restoration of the weight-bearing function of the first ray alleviates the preoperative complaint of metatarsalgia.
Metatarsophalangeal fusion leads to good outcomes also in the salvage of failed implant arthroplasty, probably due to the same reason it succeeds after failed Keller resection. Hecht et al. reported on 16 feet undergoing implant removal and fusion, 7 with autogenous tricortical graft after failed silicone arthoplasty. Two joints went on to nonunion. All patients had improved subjective pain without metatarsalgia or metatarsophalangeal or interphalangeal pain. Brodsky et al. followed 10 patients salvaged with autogenous tricortical graft for implant arthroplasty in 7, Keller resection in 1, and other procedures in the remaining 2 cases. Successful fusion was achieved in 91% of patients. These two studies show that fusion with interpositional tricortical lengthening can give good union rates and reasonable clinical outcomes.
Several potential contraindications to fusion exist. A stiff or arthritic interphalangeal joint will not compensate well in the setting of metatarsophalangeal fusion. In addition, achieving a fusion in patients with significant proximal phalangeal bone loss or those with generalized osteoporosis may be very difficult.
Allograft Versus Autograft
Few studies in the literature have assessed the use of allograft versus autograft in achieving metatarsophalangeal fusion. In fact, the only considerable attention to the subject is given by Myerson et al., who performed fusions in 24 patients followed over 5 years. A tricortical iliac autograft was used in 15 patients, medial tibial autograft in 1, and femoral allograft in 8. Fusion occurred in 79% at 13 weeks, giving a nonunion rate of 21%. Of those, none occurred in the allograft group. Although the numbers are too small to reach conclusions, it appeared that allograft might lead to better unions. The authors theorize that lack of resorption and the structural properties of the allograft conferred this advantage. However, no absolute recommendations can be made with the current studies available.
Few studies address the issue of hardware selection and its role in achieving successful fusion. Most studies use a variety of screws and plates but cannot claim superiority of one technique over another given the small number of patients involved in the series and the number of other variables present that could affect fusion rates. In a biomechanical study, Politi et al. showed that the most stable construct across the metatarsophalangeal joint was the combination of an oblique interfragmentary lag screw and dorsal neutralization plate. The plate and K-wire constructs alone were the weakest. The implications in vivo are not known.
Hecht et al. had two nonunions in their series of fusions. One occurred in the group of six patients undergoing Steinmann pin fixation, while the other occurred in the group undergoing screw and plate construct fixation. They did note, however, that the screw-and-plate group achieved union at an average of 8.25 weeks compared with 14.25 weeks in the Steinmann pin group. Myerson et al. reported nonunion in five patients undergoing fusion. Three occurred in a group of nine feet managed with plates and screws and two in nine feet managed with screws and threaded K-wires. The numbers were too small to reach conclusions, and many other technical or selection variables could also influence the rate of union beyond simply the choice of hardware. They theorized that union depends on sufficient contact and compression across the fusion site. They thought that a dorsal plate alone might compress the dorsal side of the joint and cause plantar gapping. Hardware selection depends on the bone stock available and the contour of the joint and graft. This may only be determined intraoperatively.
Management of Bone Defects
As discussed previously, although many fusion series include both patients who underwent in situ fusion and those who underwent interpositional graft, no study gives a specific number as to the amount of shortening that necessitates a lengthening (i.e., graft). Mann et al. state that, in general, shortening of the metatarsophalangeal joint of only 2 to 3 mm can potentially lead to adverse effects on the forefoot. Many studies do not routinely detail the length of shortening present or the length of graft used ( Table 40-1 ).