Arthritis of the mid- and forefoot is frequently encountered in foot and ankle practice. Primary osteoarthritis is common and the prevalence increases with age. In one epidemiological study the population prevalence of symptomatic radiographic arthritis of the mid- and forefoot in those over the age of 50 was 16.7%, with the first metatarsophalangeal joint being the most frequently involved1. The overall population prevalence of symptomatic mid/forefoot arthritis is probably lower than this, and around 4% according to most studies. This chapter offers an overview of arthritis in the mid- and forefoot and is a guide toward the various treatments.
During normal gait the movement of the foot and ankle joints is linked. Thus disease, deformity, and stiffness in one part of the foot impacts on the other parts. An understanding of the normal anatomy and biomechanics of the midfoot helps in appreciating the effects of midfoot arthritis.
The midfoot can be divided into three main longitudinal columns: medial, central, and lateral (Figure 7.1). The medial column consists of the first metatarsal and the medial cuneiform. The central column consists of the second and third metatarsals, which articulate with the middle and lateral cuneiforms. The fourth and fifth metatarsals articulate directly with the cuboid, and make up the lateral column. The less mobile columns are more stable, and vice versa. The central column has the least movement and is therefore the most stable, taking the greatest load during weight bearing with the highest contact forces. The lateral column TMTJs are the most mobile, although they only have about 10° of motion in the sagittal plane. The mobility of the lateral column is thought to be important for shock absorption and allowing the foot to accommodate to uneven ground. Overall the midfoot moves little, allowing it to act as a rigid lever arm throughout stance phase, transferring forces from the hindfoot to the forefoot.
Figure 7.1 The three longitudinal columns of the foot (red: medial; green: central; yellow: lateral).
The stability of the midfoot is achieved in part as a result of its bony architecture, and in part from the strong ligaments around the joints. The second TMTJ is highly restrained, as the base of the second metatarsal is recessed between the medial and lateral cuneiforms. In the coronal plane the cuneiforms and metatarsal bases are trapezoid in shape, with wider dorsal than plantar surfaces. Thus they act as keystones in the transverse tarsal arch (Figure 7.2). Dorsal, interosseous, and plantar ligaments support the bony structure. The intermetatarsal interosseous ligaments between the bases of the second to fifth metatarsals are the strongest of all the ligaments. The second metatarsal relies upon the Lisfranc ligament, which runs from the plantar aspect of the base of the second metatarsal to the medial cuneiform. There is no intermetatarsal ligament between the first and second metatarsals.
Figure 7.2 Line drawing of a cross-section through the level of the base of the metatarsals, showing the keystone configuration of the transverse tarsal arch.
Although the amount of movement in the midfoot joints is limited they are ideally adapted to provide rigidity at some phases of the gait cycle, while allowing suppleness and shock absorption at other phases.
Pathogenesis: Etiology, Epidemiology, and Pathophysiology
Arthritis can affect any of the midfoot joints. Arthritis of the TMTJ, particularly the second, is the most common. This is a reflection of the high loads that pass through the joint and its central role in stability. The talonavicular, naviculocuneiform, and first TMT joints develop arthritis with decreasing frequency. Arthritis in the lateral column joints is also uncommon.
Midfoot arthritis has a number of etiologies including primary osteoarthritis, post-traumatic degeneration, inflammatory arthropathy, crystal disease, and neuroarthropathy. Primary osteoarthritis is the usual cause in patients presenting over the age of 60 years2. Post-traumatic arthritis is more often seen in younger age groups, where it is usually the result of a Lisfranc injury. Lisfranc injuries can occur even after low-energy trauma and subtle injuries are often missed at initial presentation. Rarer fractures of the cuneiforms, navicular, and cuboid with articular involvement can also cause arthritis. Symmetrical polyarthropathy with widespread change across the midfoot and collapse of the medial arch is seen in rheumatoid disease. Extensive joint destruction and subluxation with a typical rocker-bottom foot deformity raises the concern of a Charcot neuroarthropathy.
The symptoms of midfoot arthritis are typically pain and swelling. Pain can be dorsal or plantar. It is mechanical in origin and is usually worse with weight bearing. Swelling is seen dorsally and can be either soft tissue, with synovitis or a ganglion, or bony osteophytes. Occasionally the deep or superficial peroneal nerve branches are irritated by prominent swellings leading to radiating pain. Typically the deep peroneal nerve is affected with pain radiating down the foot to the web space between the hallux and second toe.
As always with orthopedic examination the mantra of “look, feel, move” should be followed.
Examination begins with review of the shoes to identify abnormal wear patterns. When deformity exists it typically consists of flattening of the medial arch with midfoot abduction and supination of the forefoot on the hindfoot. In advanced cases, the plantar medial bony prominence under a deformed first TMTJ can be the cause of discomfort.
The patient is stood to assess foot position under load. Inspect the longitudinal medial arch for any midfoot collapse and look from behind to identify any heel deformity, in particular valgus. When pes planus exists the tibialis posterior tendon function should be assessed, although this may be difficult, as a single heel raise is difficult as a result of the loss of the medial longitudinal arch, and the pain, from the TMTJ arthritis.
Palpation of the foot identifies areas of tenderness, and this in itself is the best clinical indicator of the source of symptoms. A Tinel’s test may indicate nerve irritation, in particular of the deep peroneal nerve over the second TMTJ.
Movement of the midtarsal joint is limited, but deformity of the midfoot or the hindfoot should be noted. When midfoot deformity is present it is important to assess its location and to what degree it can be corrected. Tightness of the calf musculature can increase midfoot symptoms, so any ankle equinus must be identified.
At the TMTJ the individual metatarsal heads are grasped between the thumb and index finger of one hand. The corresponding TMTJ is palpated between the thumb and index finger of the other hand. Each of the TMTJs should be individually stressed in a dorsal/plantar motion and any discomfort noted. The degree of motion achieved is usually minimal – the second and third having least movement, and the fourth and fifth the most. In cases where deformity or swelling is disproportionate to pain, a Charcot neuroarthropathy should be considered. Assessment of sensibility with a 10 g Semmes–Weinstein monofilament should be undertaken routinely. Assessment of the vascular status is imperative if surgery is being considered.
Weightbearing AP and lateral foot radiographs are the workhorse of radiological diagnosis. An oblique 30° view allows assessment of all the midfoot joints. Medial and central column joints can be viewed more easily on the AP, while the oblique provides a good view of the lateral column joints. The lateral is useful to assess the second TMTJ space, which is seen parallel, but proximal, to the first TMTJ (Figure 7.3). When arthritic, narrowing of the joint can make it difficult to see on the lateral radiograph. Dorsal osteophytes are also often seen.
In cases in which there is doubt, a CT or MRI scan can be helpful in localizing the arthritic joints. Diagnostic injections with local anesthetic can also be helpful, but they should be carried out with an arthrogram, as there are numerous communications between the joints of the foot and if pain relief from the local anesthetic is used as a diagnostic criterion, it is important to know which joints have been anesthetized. If localization of pathology remains a problem, a single photon-emission computed tomography bone scan combined with CT (SPECT-CT) has been shown to help3.
A previous history of trauma with malalignment of the TMTJs on radiographs may indicate an old Lisfranc injury.
On the lateral x-ray there should be no dorsal subluxation/translation of the metatarsals. The standing lateral film allows the level of deformity to be assessed in the planus foot. This is important for surgical planning. Meary’s line, a line drawn longitudinally down the midaxis of the talar neck, should be collinear with the midaxial line of the first metatarsal. Where deformity exists the lines are not parallel and their intersection point indicates the apex of the deformity (Figure 7.4). The apex may be at the TMTJ or further proximally at the naviculocuneiform or talonavicular level.
The mainstay of treatment is simple analgesic medication and in-shoe orthoses with a medial-arch support. Weight loss also helps to reduce load. Occasionally an AFO with a shoe rocker may be of help, but many patients are reluctant to accept this. Where dorsal osteophytes are the main issue, alternative lacing patterns on shoes can relieve pressure. Cut outs, stretching of the shoe upper, or padding over the area may also be helpful.
Operative intervention should be reserved for cases that fail conservative management over a prolonged period, usually a minimum of six months. The surgical procedure depends upon the cause of the discomfort. It may simply involve removal of the dorsal osteophyte with release of the deep peroneal nerve.
More commonly, the pain arises from the joint. Arthrodesis is the mainstay of surgical treatment. As discussed above, surgical planning starts with determination of which joints are involved. This is usually based upon a combination of clinical examination and plain radiographs, although diagnostic local anesthetic injections with an arthrogram, MRI scanning, and SPECT-CT may be helpful.
The basis of all fusions includes joint preparation to bleeding bone and stable fixation. It is important to consider whether the fusion is in situ, or requires deformity correction. Even a fusion in situ requires care, for example avoiding excessive shortening or dorsiflexion in a TMTJ fusion.
The workhorse for surgery in the midfoot is a TMTJ fusion; however, not all TMTJ fusions are in situ. Deformity correction may include fusion of the foot with a rocker bottom, correction of a hallux valgus (Lapidus procedure), and TMTJ fusion as part of a flat-foot correction. It has been shown that if deformity is present, correction of that deformity in association with the TMTJ fusion gives better results than simple in situ fusion4. The aim is to achieve a stable, pain-free, plantigrade foot without bony prominences, which fits comfortably into standard shoes. Deformity correction should proceed from proximal to distal to ensure that the foot is plantigrade at the end of surgery, thus tendo Achillis lengthening and correction of hindfoot alignment should be addressed first. It is not until a neutral hindfoot position has been achieved that the midfoot is positioned.
The technique of TMTJ fusion varies depending on the indication. The first ray can either be approached through a dorsal or medio-plantar approach. In all cases the joint is denuded of cartilage and taken back to bleeding bone. Minimal bone resection is usually required, even if deformity correction is required. This prevents shortening of the first ray and the consequent development of transfer metatarsalgia in the lesser rays. Thus with a Lapidus procedure only the most minimal laterally based wedge is excised, and even then the first metatarsal is translated plantarward and held in place with a stepped plate to hold the ray in this position.
One of the commoner deformities is of TMTJ collapse with abduction; this is often associated with a rocker-bottom type deformity. The patient complains of prominence and callosity under the midfoot. A corrective first TMTJ fusion with a closing plantar-medial wedge is required. This is best undertaken through a plantar-medial approach, which allows plantar plating of the joint (Figure 7.5). Cadaveric studies show plantar plating to be more biomechanically stable than dorsomedial plating, with greater initial and final stiffness and a greater load to failure5.
The lesser TMTJs are approached through a dorsal incision, taking care not to damage the deep peroneal nerve and the dorsalis pedis artery. When a plantar medial approach is used to reach the first TMTJ a second dorsal incision can be made slightly more laterally between the second and third TMTJs to access these joints. This leaves a good skin bridge and the incision is not directly over the neurovascular bundle. Joint preparation needs to be undertaken with care. As the TMTJs are approached through a dorsal incision there is a tendency to remove more bone dorsally than plantarly, which can lead to fusion in a dorsiflexed position. In many cases, lesser metatarsal length will remain unaltered, in these cases one approach is to leave the plantar cortex intact. This creates a dorsal “trough,” which can be filled with local bone graft from the os calcis.
A variety of internal fixation methods have been described including staples, screws, and locked and non-locked plates. The approach and degree of fixation are dependent on the combination of joints involved and the amount of bone loss or deformity that coexists. An isolated in situ fusion can be adequately stabilized with screws, although plating, particularly of the first ray, is commonly undertaken. If a screw is used from distal to proximal, countersinking the metatarsal shaft with a burr is useful to prevent the screw head pushing upward and cracking the dorsal cortex – so-called Manoli holes (Figure 7.6).
Figure 7.6 Manoli holes: a recessed countersink is burred from the metatarsal cortex to accommodate the head of the screw. If this is not undertaken there is a risk of the dorsal cortex of the metatarsal fracturing, as a result of upward pressure from the screw.
Not every case requires a bone graft. Nevertheless, even if it is required, the amount is often small and can be taken locally from the os calcis or proximal tibia. Graft from the calcaneus is easy through a short lateral incision. Proximal tibial bone can be accessed through a lateral cortical window. Complications, such as sural nerve injury, are rare although up to 13.8% of patients report some residual symptoms along the lateral border of the calcaneus6. One study that looked at pain outcome scores across multiple sites for autologous bone-graft harvest showed 12% of subjects reported clinically significant pain at the harvest site at 24 weeks and 8.5% at 52 weeks postoperatively, with each of the lower extremity harvest sites (calcaneum, distal tibia, and proximal tibia) having greater rates of persistent pain at one year than the iliac crest bone-grafting group7. It is therefore important to remember that bone grafting is not without morbidity, and to consider this both in surgical planning and when obtaining consent.
Results and Complications
When restricted to patients who have failed conservative therapy for degenerative and inflammatory arthropathy, long-term overall operative results of arthrodesis of the midtarsal and TMTJs have shown 93% satisfaction at six-year follow-up2. Union rates between 92 and 98% for primary fusions are quoted2, 8–9. A study of 72 patients showed that union occurred in the majority, 74%, by nine weeks – although 4% of patients took up to 16 weeks9.
The complication rates vary from 4 to 17%8–9. The most frequent complication reported, in 13%, is malunion – with pain and prominence under the lesser metatarsal heads2. This does not always require further operative intervention, although Komenda found the need to perform a dorsal closing metatarsal wedge osteotomy to correct this in 2 out of 32 patients10. In the first ray, pain under the sesamoids is also reported as a result of the lack of flexibility4. Wound-healing problems, superficial infection, nerve injury with painful neuroma formation, stress fractures, and chronic regional pain syndrome are also reported2, 10.
Midfoot fusions for Charcot neuropathic deformity are more challenging and prone to more complications than a standard midfoot fusion. The aim is to achieve a stable, plantigrade, shoeable foot, which is stable (Figure 7.7). In patients with diabetes mellitus this can be satisfactorily achieved in about 60% non-operatively. The remaining patients are at increased risk of amputation, although Pinzur detailed a 92% salvage rate with correction and ring fixation in this group11. There are a number of factors that need to be taken into consideration in operating upon these patients, these include the blood supply, ulceration and infection, the magnitude of the deformity, techniques of fixation, and the length of immobilization.
Ulceration and infection may necessitate that surgery is staged, with primary debridement and exostectomy to heal the ulcer, followed by reconstruction with arthrodesis. External, rather than internal, fixation may also be chosen.
The magnitude of the deformity is often extreme, requiring significant bony resection. It is important at the end of surgery to be left with a relaxed soft tissue envelope, which can be easily closed.
Fixation techniques will need to vary from the standard. In the presence of persisting ulceration and infection, a ring fixator may be chosen. Internal fixation has to be robust. Multiple axially placed intramedullary screws have been proposed as providing a stable construct to achieve and maintain correction of the deformity12. However, the insertion of intramedullary screws is technically demanding, and failures have been reported13. Furthermore, when tested biomechanically, plantar plating and intramedullary screw fixation show equal stiffness and load to failure with no notable biomechanical difference between the techniques14.
The duration of immobilization needs to be increased to reflect that, on average, complete osseous union takes six months in the neuropath. Consequently, protection and weightbearing limitations need to be prolonged. The old adage, of taking the standard time for treatment and doubling it, is a start.
The fourth and fifth TMTJs are the most mobile, and symptomatic arthritis is rare. Unsurprisingly, the literature is sparse with only small case series. It is usually advised that arthrodesis of these joints should be avoided and that fourth and fifth TMTJ arthritis should be treated with excisional arthroplasty. Raikin reviewed 28 patients undergoing midfoot arthrodesis: six had an isolated lateral column fusion and the others had a combined medial, central, and lateral-column arthrodesis. At two-year follow-up the fusion rate was 92% with pain levels improving from 8.2 to 2.4 in those with isolated lateral column disease. Thus despite the concern that arthrodesis of the lateral column is ineffective it has been reported with good results15.
Excisional arthroplasty with tendon interposition has also been shown to be effective. In a small series of 12 patients, three-quarters were satisfied with the result of surgery16. In an attempt to maintain movement of the fourth and fifth TMTJs, interpositional arthroplasty with ceramic components has also been tried. Despite reported improvement in pain levels the study sizes are very small and the follow-up less than three years17–18.
Isolated talonavicular fusion has a 93% satisfaction rate and is most commonly carried out for adult-acquired flat foot, isolated osteoarthritis, or rheumatoid arthritis. It should be borne in mind that isolated fusion of the talonavicular joint almost completely abolishes subtalar movement; nevertheless, it is less invasive than a triple fusion and is therefore not uncommonly undertaken in patients with isolated TNJ disease19.
Arthrodesis of the naviculocuneiform joints may form part of a medial-column stabilization, although isolated fusion of this joint complex is uncommon. Isolated fusion may be indicated in a small group of patients with a degenerate naviculocuneiform joint (Figure 7.8) or as part of a planovalgus foot correction. One retrospective analysis of 33 fusions, undertaken for a mixture of degeneration and planovalgus deformity, showed it to be safe and reliable with 97% union rates. Union did, however, take slightly longer than fusions elsewhere in the foot with an average of 22 weeks. The patient who had a non-union was revised and subsequently went on to have a successful union20.
The normal first MTPJ is a cam-shaped, hinge joint with a greater range of dorsiflexion (40–100°) than plantar flexion (3–43°). The maximum standing range of movement on tiptoe stance averages 65°, but the functional range of motion during walking is only 38°. In running a range of 60° may be needed.
Hallux rigidus is commoner in women and its prevalence increases with age. In the over 50s, there is a 7.8% prevalence of symptomatic first MTPJ arthritis. Bilateral disease is common, affecting up to 80% of patients, although the severity of symptoms differs on each side21.
The majority of cases of hallux rigidus are idiopathic, although it can be secondary to trauma, inflammatory arthropathy, or crystal arthropathy. There have been many theories as to the etiology of idiopathic hallux rigidus with a number of authors looking at anatomical variations as the underlying mechanical factor leading to disease. Metatarsus primus elevatus, metatarsus primus varus, a flattened first metatarsal head, hypermobility, increased first metatarsal length, and calf tightness have all been implicated, but none has been proven to be the definitive cause.
With metatarsus primus elevatus the elevation may be idiopathic, secondary to malunion of a fracture, or iatrogenic, following a first metatarsal osteotomy. It is postulated that elevation of the first metatarsal head leads to the proximal phalanx being relatively plantarly subluxed. Therefore during early toe-off the base of the proximal phalanx impinges against the dorsal metatarsal head, causing osteochondral damage of the dorsal aspect of the metatarsal head22. This fits with the pattern of wear that is seen in hallux rigidus with early loss of the dorsal articular cartilage. Opposing the view that elevatus is etiological is evidence that the elevatus is a secondary change in joints that are already degenerative. Furthermore a study by Meyer has shown that the elevated first metatarsal is a normal finding in midstance23.
A number of grading systems for hallux rigidus exist. In general, they assess both the range of movement and the radiographic features of the joint. The most widely used classification is that described by Hattrup and Johnson24 (Table 7.1), which has subsequently been modified by Coughlin25 (Table 7.2).