THE CLINICAL AND RADIOGRAPHIC OUTCOME OF REVISION SURGERY FOR FAILED TRIPLE ARTHRODESIS
March 18th, 1997
Steven L. Haddad, MD; Mark S. Myerson, MD;
Richard F. Pell IV, MD; and Lew C. Schon, MD
Between 1987 and 1994, we treated 33 patients with surgical revision for failed triple arthrodesis, 28 (29 feet) of whom returned for final examination(mean, 4.4 years; range, 2 to 7 years). The average age of these 16 women and 12 men was 46 years (range, 14 to 69 years). Before the revision procedure, patients had undergone nonoperative therapies for an average of 3.7 years (range, 0.5 to 12 years) and an average of three foot operations (range, 1 to 6) after the primary triple arthrodesis.
All patients were managed with rigid internal fixation via cannulated screws and power staples. Calcaneal osteotomy and/or revision of the transverse tarsal arthrodesis via appropriate saw cuts and bone wedges were used. Iliac crest bone graft was added, when a bone block arthrodesis was required, for those patients with nonunion or ankle impingement.
Arthrodesis was achieved in all 29 feet, although 4 patients (4 feet) (14%) required additional procedures for malunion (2), deformity recurrence (1), deep infection (1), and skin graft (1). Comparison of pre- and postoperative radiographs showed an average correction of 13o (p = 0.069) in all planes. Comparison of the average pre- (retrospective) and postoperative American Orthopaedic Foot and Ankle Society 94-point hindfoot and ankle scores showed a significant improvement: 31 points (range, 13 to 61 points) versus 59 points (range, 24 to 91 points), respectively (p < 0.05). On a scale of 0 to 10 points, average patient satisfaction was 7.8 points (range, 2 to 10 points).
This study demonstrated a satisfactory improvement in patient outcome after surgical correction of failed triple arthrodesis. We conclude that such a revision, although rather complex, may be attempted to establish a plantigrade foot free of infection and shoeable without an orthosis or brace.
Triple arthrodesis is a commonly performed procedure for stabilizing and correcting a variety of pathologic processes affecting the foot. Ryerson15 is given credit for first describing the triple arthrodesis in 1923 after adding lateral column fusion to assist in correcting severe varus deformities. Early procedures relied on molded plaster casts to maintain correction, but malunion and recurrence of the deformity were common.18 The addition of internal fixation decreased some of these complications, but introduced a new era of technical errors.
The literature is replete with long-term studies of primary triple arthrodesis,1,4,7,8,18-21,23,24 evaluating this treatment for conditions such as Charcot-Marie-Tooth disease and other neuromuscular disorders,10,23,24 trauma,3,12 and more recently, flatfoot deformity.13,16 All such studies report poor results with deformity recurrence rates of 9 to 20%,4,8,18,24 nonunion rates of 6 to 33%,5,10,18,24 and incomplete correction of the deformity.2,18,23 No study adequately addresses revision of failed triple arthrodesis (persistent pain and mechanical deformity).
Failure of these triple arthrodesis procedures may be attributable to worsening arthritis of adjacent joints.6,11,14,18 We have focused our study on the failure of the initial arthrodesis procedure to achieve its overall objective and have developed a systematic, algorithmic approach for the correction of symptomatic residual or recurrent deformity. The outcome of this approach has been assessed clinically and radiographically.
Materials and Methods
Between August 1987 and February 1994, 38 patients underwent revision of a previously performed triple arthrodesis. Patients with a diagnosis of neuropathic arthropathy (N = 5) were excluded from the study. Thirty-three patients were thus eligible for the study. Two patients died from causes unrelated to this surgery, two patients refused follow-up, and one patient was lost to follow-up examination. The remaining 28 patients (29 feet, 12 left, 17 right) returned to our institution for a follow-up examination at an average of 4.4 years (range, 2 to 7 years) after revision surgery. At the time of revision surgery, the 16 women and 12 men averaged 46 years in age (range, 14 to 69 years), 176 pounds in weight (range, 90 to 285 pounds), and 67 inches in height (range, 62 to 76 inches). Patients stated that their pain level before revision surgery averaged 5.8 points (range, 2 to 10 points) on a scale of 0 (no pain) to 10 (pain so great the patient desires leg amputation). Fifteen of the patients were smokers, consuming an average of .9 packs of cigarettes/day. Preexisting medical conditions included psoriasis (1), lupus (1), heart disease (3), and gout (1). We recognize that retrospective scoring is subject to patient recollection, but we have included the pre- and postoperative AOFAS scores for purposes of comparison.
The primary triple arthrodeses had been performed for trauma (11 feet: 7 calcaneal fractures, 3 talar fractures, 1 sciatic nerve injury occurring at birth), hereditary sensorimotor neuropathy (4 feet), rheumatoid arthritis (7 feet), congenital deformity (4 feet: 1 residual clubfoot, 3 tarsal coalitions), adult acquired flatfoot deformity (2 feet), and poliomyelitis (1 foot). Five patients (5 feet) had been injured on the job and were receiving benefits through workmanís compensation. The failed primary arthrodeses resulted in multiplanar deformity (eg equinovarus or equinovarus with rocker-bottom) (10 feet), varus hindfoot malunion (8 feet), valgus hindfoot malunion (5 feet), nonunion (4 feet), and rocker-bottom deformity (2 feet). The revision procedure was performed after long-term nonoperative therapies (average, 3.7 years; range, 0.5 to 12 years) and other foot operations (average, 3; range, 1 to 6) proved unsuccessful.
Our surgical approach to revising a failed triple arthrodesis followed an algorithm that sequentially addressed each symptomatic fixed deformity (Fig. 1).
General anesthesia was used for 14 patients, regional ankle block for 12 patients, and spinal anesthesia for 2 patients. A tourniquet was used for 12 patients. Concerns about wound closure after revision surgery determined the incision placement, especially with regard to preoperative hindfoot valgus deformities, for which lateral wound closure is complicated. Previous surgical incisions were used only if they permitted adequate exposure and closure without wound tension. If it was determined that the preexisting incisions might lead to excessive soft-tissue traction or tension, the incision was planned along or parallel to the longitudinal axis of the foot. The osteotomies were determined by the plane of the deformity.
Multiple deformities were often present, mandating a stepwise approach to surgical correction (Fig. 1), ie hindfoot varus/valgus was addressed before forefoot supination/pronation. The shape of the foot was carefully assessed, and the apex of the deformity was determined. For example, for a varus-supination deformity, the apex often corresponded to the cuboid or the fifth metatarsal base; recognizing the multiple components of deformity, we attempted to obtain maximum correction proximal to the apex of the deformity.
Hindfoot varus malunion was addressed by one of two methods: a closing wedge osteotomy through the previous fusion site at the level of the subtalar joint (Fig. 2) or a lateral closing wedge osteotomy with lateral translation through the tuberosity of the calcaneus (Fig. 3). If a callus was present under the fifth metatarsal head or base, fixed forefoot supination compounded the deformity. Thus, a calcaneal osteotomy alone would not be sufficient (see below). Residual hindfoot valgus deformity was corrected with a medial displacement osteotomy through the tuberosity of the calcaneus. No attempt was made to tilt the tuberosity into varus or to remove a medially based wedge. The incision was identical for both varus and valgus correction. An oblique incision was made, beginning two finger widths anterior to the insertion of the Achilles tendon and 1 cm inferior to the tip of the fibula. Cuts and bone wedges were made with an oscillating saw. The calcaneal tuberosity was translated 1.0 to 1.5 cm in the appropriate direction and secured with a cannulated cancellous lag screw.
For residual forefoot supination and pronation, medial and lateral incisions were used. Medially, the incision was located in the interval between the anterior and posterior tibialis tendons. Laterally, the incision began at the tip of the fibula, crossed the dorsal lateral aspect of the calcaneocuboid joint, and ended over the cuboid. As noted above, a straight (not curved) incision was used to avoid problems with wound closure. Once these dissections were carried to bone, a periosteal elevator was used to strip the soft tissues off the bone across both the dorsum and plantar surfaces of the transverse tarsal joint fusion mass. Malleable retractors were placed to protect the soft tissues (both dorsal and plantar). After fluoroscopic confirmation of the planned osteotomy, an oscillating saw with a long blade was used to make a transverse cut across the foot at the level of the former calcaneocuboid and talonavicular joints, preserving as much of the cuboid and navicular bone as possible to facilitate fixation. The distal portion of the foot was then rotated into a plantigrade, neutral position (Fig. 4). Guide pins were placed across the osteotomy site and large cannulated screws were used for fixation. If cannulated screws were not adaptable to the plane of correction (particularly laterally), power staples (3M, St. Paul, MN) were used to stabilize the transverse tarsal joints.
To correct an abducted or adducted forefoot, a biplanar closing wedge osteotomy was performed through the fusion mass (medial for excessive abduction, lateral for excessive adduction). Simple rotation of the fusion through the osteotomy was not sufficient to correct this deformity, as it was for pronation or supination. To determine the angle of the wedge, a proximal guide pin was placed as perpendicular as possible to the plane of the hindfoot. A second distal guide pin was placed as perpendicular as possible to the plane of the forefoot (Fig. 5). Saw cuts were made parallel to these two pins, taking care to preserve the cuboid and navicular for distal fixation. Cannulated screws or power staples were used to stabilize these deformities.
A rocker-bottom deformity was corrected with a plantar-based closing wedge osteotomy, with malleable retractors to protect the dorsal and plantar structures and an oscillating saw positioned as described above. The prominent mass of bone was resected within this wedge (Fig. 6).
Multiplanar deformity was approached using all of the above-mentioned principles. Guide pins were used to establish angles for the biplanar wedge of bone to be removed. In most cases, these deformities required removal of multiple bone wedges, the position of each one dependent on the plane of the deformity.
Generally, bone graft supplementation was unnecessary with the above-mentioned osteotomies. In our patient group, iliac crest autograft was used for the four patients (#11, 12, 13, 14) with nonunion of the primary triple arthrodesis and for one patient (#1) with anterior ankle impingement who required a talocalcaneal bone block arthrodesis. The remaining 24 patients either required no additional bone graft or were managed with local bone graft obtained from the wedges of removed bone.
All patients were managed initially with a non-weight-bearing compressive posterior plaster splint for 10 days, followed by a non-weight-bearing short-leg cast for an additional 4 weeks. Weight-bearing was begun in either a walking cast or a range-of-motion walker boot locked in a neutral position for another 4 to 6 weeks, until union was evident. Time to arthrodesis was determined both clinically (stability at the osteotomy site) and radiographically (trabeculation across the osteotomy site).
Subjective assessment. Before the final follow-up examination, each of the 28 patients was asked to complete a questionnaire that included the subjective portion of the American Orthopaedic Foot and Ankle Society (AOFAS) ankle-hindfoot scale (60 points)9 and some additional information. Patients were asked if they were satisfied, to rate their satisfaction on a scale of 0 (not satisfied at all) to 10 (completely satisfied) points, and whether they would have the revision surgery performed again.
Clinical assessment. All 28 patients returned to our institution for clinical follow-up, including determination of gait abnormality, sagittal motion, ankle-hindfoot stability, and alignment via the AOFAS Ankle-Hindfoot scale (34 points)9 and a physical examination. During the physical examination, the presence of ankle instability and the presence, location, and symptoms of calluses were assessed and recorded. Knee varus, valgus, and recurvatum, ankle dorsiflexion, and ankle plantarflexion measurements were then made with the patient standing, and the patient's ability to perform single and bilateral heel rise was tested.
Radiographic analyses. During the follow-up visit, we obtained weight-bearing anteroposterior (AP) and lateral radiographs of both feet and ankles, an additional "marker" AP radiograph (with strips of malleable lead-free solder [Safe-Flo Silver Lead-Free Solder, Oatey, Cleveland, OH] in place on the patientís hindfoot and ankle malleoli) to assess true varus or valgus of the heel (Fig. 7), and stress views (Telos Stress Device, Austin & Associates, Inc., Fallston, MD) of both ankles to analyze anterior drawer and varus/valgus instability in each patient.
Measurements on these postoperative radiographs were then obtained and compared to those on matching preoperative radiographs. On the AP radiograph, the talus-first metatarsal and talonavicular coverage angles17 were measured. On the lateral radiograph, the talus-first metatarsal, talocalcaneal, and calcaneal pitch angles were measured, as was the height of the base of the first and fifth metatarsals to the floor.
The marker AP radiograph assessed hindfoot varus or valgus. A line was drawn perpendicular to the axis of the heel by determining two midpoints between the medial and lateral arms of the marker around the heel. A second line was drawn parallel to the long axis of the tibia. The angle of intersection of these lines determined the amount of varus or valgus of the hindfoot.
Ankle instability was evaluated on the lateral stress view by creating a line perpendicular to the talus, connecting to the most posterior-distal portion of the tibia, and measuring the distance. On the AP stress view, we measured medial and lateral instability by drawing lines parallel to the articular surfaces of the tibia and talus and determining the angle of intersection.
The statistical tests used to evaluate the data included Studentís t-test, paired t-test, one-way ANOVA (analysis of variance with a Scheffe and Tukey comparison), and Pearsonís correlations using the SPSS/PC (SPSS Inc., Chicago, IL) statistical software program.
Comparative analyses were performed on the radiographic data to assess improvement in the lateral talocalcaneal, talus-first metatarsal, and calcaneal pitch angles for hindfoot varus and valgus pre- and postoperatively. Data for the operative and nonoperative foot of each patient (except the patient with bilateral procedures) were compared in a similar fashion. Although this was a retrospective study, pre- and postoperative AOFAS ankle-hindfoot scores were compared. (We do, however, recognize the limitation of patient recall and subjective exaggeration of preoperative pain and dysfunction.) In addition, analyses were performed to define 1) any correlation between deviation from an ideal hindfoot position (0 to 5o of valgus) and the postoperative AOFAS score and 2) any correlation between the magnitude of correction in hindfoot varus/valgus (in degrees) and the improvement in the AOFAS score. This latter correlation was performed to determine if shifting the heel into a more ideal position would improve the patientís outcome as manifested by the AOFAS score, regardless of the heel's final position.
Table 2 shows questionnaire response averages. In addition, patients were asked to rate their satisfaction with the procedure on a 0 (completely unsatisfied) to 10 (completely satisfied) scale; the average overall satisfaction was 7.8/10 points. All patients stated that they would have the revision procedure performed again.
On physical examination, we found two patients (#13, 18) with symptomatic callouses, both located under the first metatarsal head in patients with some residual forefoot pronation. Ankle motion averaged 10o of dorsiflexion and 28o of plantarflexion. Preoperatively, no patient could perform a single heel rise; postoperatively, 6 patients (#3, 5, 10, 14, 22, 15) (21%) could perform this exercise. Three patients (#12, 14, 20)had clinically significant ankle instability, verified radiographically, and two other patients (#7, 16)had instability manifested only radiographically.
Comparison of the average pre- and postoperative AOFAS ankle and hindfoot scores (59 versus 31 points, respectively) showed a highly significant (p < 0.01) average improvement of 32 points. Comparison of pre- and postoperative shoe wear modifications also demonstrated a significant improvement (p = 0.01).
Overall hindfoot alignment (including preoperative varus and valgus deformities) averaged neutral (0o) in the surgically corrected feet, as compared to an average of 2o valgus in the nonoperative feet.
The comparisons of average preoperative, postoperative, and nonoperative radiographic measurement data for hindfoot varus and valgus are presented in Table 5. Data comparing the changes in heights of the first and fifth metatarsals are presented in Table 6. Abduction was improved an average of 21o as measured on the AP talar-first metatarsal radiographs. Rocker-bottom deformity correction demonstrated an average of 27 and 20 mm improvement in the height of the first and fifth metatarsal bases, respectively. In addition, the lateral talar-first metatarsal angle improved an average of 17o for rocker-bottom deformity. Comparative analyses of these data revealed that the changes were not statistically significant. However, there was a weakly positive correlation (r = .2403) between the improvement of preoperative hindfoot varus/valgus deformity and the improvement in the patient's AOFAS score. As expected, there was a weakly negative correlation (r = - .2078) between the postsurgical difference from the ideal hindfoot valgus (0 to 5o) and the postoperative AOFAS score.
Time to Fusion
Arthrodesis occurred rapidly, as determined both clinically and radiographically. The average time to fusion was 8.9 weeks (range, 4 to 12 weeks). Although 4 weeks may not appear credicle, and realistically arthrodesis may not have occurred, we used the same subjective and objective parameters to evaluate the presence of union. This was determined by the absence of pain either with foot manipulation or with weight-bearing, minimal swelling, no palpable warmth, and confirmatory radiographic findings.
Four patients (14%) sustained major complications, defined as a condition necessitating additional surgery, and seven patients (25%) had minor complications.
Many studies have addressed the long-term results of primary triple arthrodesis. Wetmore and Drennan23 reviewed 30 operative feet of patients undergoing triple arthrodesis for Charcot-Marie-Tooth disease. The average follow-up was 21 years and there were poor results in 47%. Residual cavovarus deformity as determined on immediate postoperative radiographs was seen in 30% of the patients, and an additional 23% developed recurrent cavovarus deformity. Forty percent of these patients experienced moderate or severe pain, with painful calluses, metatarsalgia, and ankle arthritis.
Graves et al7 reviewed the long-term results of triple arthrodesis performed for 17 patients (18 feet) more than 50 years old (average age, 66 years). Contrary to the study by Wetmore and Drennan,23 most of these patients were affected by nonneuromuscular disorders such as posterior tibial tendon rupture and rheumatoid arthritis. Although deformity did not recur in these patients, subsequent degenerative changes in the ankle and joints distal to the arthrodesis and three nonunions led to persistent discomfort in 65%; 41% had no improvement in the distance they could walk. Yet Graves et al7 found no association between postoperative forefoot position or postoperative correction (as determined radiographically) and patient satisfaction.
Sangeorzan and Hansen18 examined 44 patients 5 years after triple arthrodesis. This group was younger (average age, 41 years) than that in the series of Graves et al7 and had a wide range of primary diagnoses, including 11 patients with Charcot-Marie-Tooth disease. All surgical procedures involved rigid internal fixation with screws. Using strict subjective, objective, and radiographic criteria, 77% demonstrated good results, with a 9% failure rate (pseudoarthrosis, 2; residual hindfoot varus, 2). Radiographically, all patients demonstrated significant improvement between pre- and postoperative measurements, with similar correction in patients presenting with varus and valgus deformities. The authors found that the average patient complained of mild pain at least once per month (with little or no restriction in daily activities) and that there was no recurrence of deformity.
These three studies highlight the fact that triple arthrodesis often fails to resolve the problems for which it was undertaken; in fact, it may produce additional problems through overcorrection of deformity secondary to rigid internal fixation, leaving many patients dissatisfied with the outcome of their surgery. Although these unsatisfactory results are documented in these and many other studies, no report in the literature adequately addresses how to analyze and approach the failed triple arthrodesis.
In the only other report in the literature discussing revision of failed triple arthrodesis, Stephens and Saleh22 performed a crescentic calcaneal dome osteotomy in five patients to revise symptomatic deformity after triple arthrodesis. Although they purported that this osteotomy allows for correction of many different types of residual forefoot deformities, the proximal location of the osteotomy makes this procedure less utilitarian than multiplanar reconstruction as described in the current study. Unfortunately, the study group was limited, making it difficult to extrapolate conclusions about the usefulness of this technique to the many deformities present in patients with failed triple arthrodesis.
In the current series, deformity occurred due to one of three technical reasons. First, the deformity was not completely corrected at the time of surgery due to underappreciation of the amount of correction required and appropriate position of the foot. Second, recurrence of a previous deformity may occur. This is most obvious in patients with a progressive neuromuscular imbalance (ie type 1 Charcot-Marie-Tooth disease). Third, the surgeon may cognitively understand the position in which the foot should be fused but, for some reason, may be unable to achieve this technically.
We have outlined a systematic approach to correcting residual deformities, starting at the most proximal aspect of the deformity and proceeding proximal to distal. When approaching the malpositioned triple arthrodesis, a standard technique is not used, since correction depends on the presence and magnitude of deformity in the hindfoot, midfoot, and forefoot. Hindfoot deformities are corrected with an osteotomy of the posterior tuberosity of the calcaneus, a valgus heel requires a medial translational osteotomy, and a varus heel may be addressed with a lateral translational osteotomy or by a laterally based closing wedge through the subtalar fusion mass. Occasionally, a closing wedge and translational osteotomy is needed through the tuberosity (Figs. 2 and 3).
If the forefoot is supinated or pronated, a transverse osteotomy is performed through the calcaneocuboid and talonavicular fusion sites. This derotational transverse osteotomy allows the surgeon to place the forefoot into proper position. If abduction or adduction is a component of the deformity, it can be addressed by a wedge of bone taken through the transverse tarsal fusion site (Figs. 8 and 9).
If a rocker-bottom deformity exists at the calcaneocuboid and talonavicular joints, a plantar closing wedge through the transverse tarsal fusion site is used (Fig. 10). A combination of wedging and derotation may be necessary for complex multiplanar malunions. For example, a pronated, abducted forefoot will require a derotational transverse osteotomy with an additional medial closing wedge at this site.
Complications were primarily due to wound problems. Three patients with infection (two superficial, one deep) and one patient with wound dehiscence are not unexpected in light of the potential tension placed on wounds after correction of substantial deformity, particularly with lateral wounds after correction of valgus malunion. Care must be taken to use longitudinal incisions whenever possible and avoid preexisting incisions that would place undue tension after wound closure.
Considering the poor preoperative condition of this group of patients, our surgical approach achieved excellent success, as evidenced by substantial improvement in AOFAS scores, alignment, pain, and varus/valgus and supination/pronation deformities. The 14% major complication (reoperation) rate was not unexpected in such a complex group of patients.
Our surgical goal was to convert a rigid, unbalanced foot that often required supplemental orthoses for weight-bearing into a plantigrade foot capable of fitting into a regular shoe without an orthosis or brace. Tables 1 and 4 demonstrate the success of our techniques. The difference between these groups before and after revision was statistically significant (p = 0.01). This improvement from restrictive to nonrestrictive shoe wear modifications would have an additional benefit of decreased maintenance costs.
The success of this corrective surgery may also be documented by comparison of pre- and postoperative pain. Before revision surgery, the average patient score was 5.8 on a scale where 10 equaled pain great enough to prompt amputation. After revision surgery, the average pain score was 2.1 (AOFAS scale of 1 to 4), equivalent to mild, occasional pain.
Most importantly, all patients stated they would undergo the revision procedure again. Generally, these patients have a substantial level of pain from triple arthrodesis malunion. Prolonged conservative management (average, 3.7 years) offered no long-term relief for our patients' symptoms. This underscores the point that primary triple arthrodesis must be performed with meticulous attention to detail to achieve a plantigrade, asymptomatic foot. When patients are unsatisfied with fusion of their hindfoot, physical examination and radiographic analysis should be used to determine the deformity causing their symptoms.
Our surgical algorithm (Fig. 1) is an appropriate method for addressing such residual or recurrent deformities, as demonstrated by the improved outcome of our patient group. We conclude that such a revision, although rather complex, can be attempted to establish a plantigrade foot that is free of infection and capable of fitting in a regular shoe without an orthosis or brace.
1. Angus, P.D., and Cowell, H.R. Triple arthrodesis. A critical long-term review. J. Bone Joint Surg. 68B:260-265, 1986.
2. Bennett, G.L., Graham, C.E., and Mauldin, D.M. Triple arthrodesis in adults. Foot Ankle, 12:138-143, 1991.
3. Connolly, J.F. Persistent heel pain twenty years after calcaneal fracture and triple arthrodesis relieved by lateral decompression. J. Trauma, 27:809-810, 1987.
4. Figgie, M.P., O'Malley, M.J., Ranawat, C., Inglis, A.E., and Sculco, T.P. Triple arthrodesis in rheumatoid arthritis. Clin. Orthop. 292:250-254, 1993.
5. Galindo, M.J., Jr., Siff, S.J., Butler, J.E., and Cain, T.E. Triple arthrodesis in young children: a salvage procedure after failed releases in severely affected feet. Foot Ankle, 7:319-325, 1987.
6. Gelberman, R.H., Woo, S.L., Lothringer, K., Akeson, W.H., and Amiel, D. Effects of early intermittent passive mobilization on healing canine flexor tendons. J. Hand Surg. 7:170-175, 1982.
7. Graves, S.C., Mann, R.A., and Graves, K.O. Triple arthrodesis in older adults. Results after long-term follow-up. J. Bone Joint Surg. 75A:355-362, 1993.
8. Haritidis, J.H., Kirkos, J.M., Provellegios, S.M., and Zachos, A.D. Long-term results of triple arthrodesis: 42 cases followed for 25 years. Foot Ankle Int. 15:548-551, 1994.
9. Kitaoka, H.B., Alexander, I.J., Adelaar, R.S., Nunley, J.A., Myerson, M.S., and Sanders, M. Clinical rating systems for the ankle-hindfoot, midfoot, hallux, and lesser toes. Foot Ankle Int. 15:349-353, 1994.
10. Mulier, E., De Rijcke, J., Fabry, G., and Mulier, J.C. Triple arthrodesis in neuromuscular disorders. Acta Orthop. Belg. 56:557-561, 1990.
11. Myerson, M. Arthrodesis for diabetic neuroarthropathy. In Current Therapy in Foot and Ankle Surgery, Myerson, M. (ed.). St. Louis, Mosby-Year Book Inc, 1993, pp. 116-122.
12. Myerson, M., and Quill, G.E., Jr. Late complications of fractures of the calcaneus. J. Bone Joint Surg. 75A:331-341, 1993.
13. Myerson, M.S. Acquired flatfoot in the adult. Adv. Orthop. Surg. 22:155-165, 1989.
14. Papa, J., Myerson, M., and Girard, P. Salvage, with arthrodesis, in intractable diabetic neuropathic arthropathy of the foot and ankle. J. Bone Joint Surg. 75A:1056-1066, 1993.
15. Ryerson, E.W. Arthrodesing operations of the feet. J. Bone Joint Surg. 5:453-471, 1923.
16. Sammarco, G.J. Technique of triple arthrodesis in treatment of symptomatic pes planus. Orthopedics, 11:1607-1610, 1988.
17. Sangeorzan, B.J., Mosca, V., and Hansen, S.T. Effect of calcaneal lengthening on relationships among the hindfoot, midfoot, and forefoot. Foot Ankle, 14:136-141, 1993.
18. Sangeorzan, B.J., Smith, D., Veith, R., and Hansen, S.T., Jr. Triple arthrodesis using internal fixation in treatment of adult foot disorders. Clin. Orthop. 294:299-307, 1993.
19. Santavirta, S., Turunen, V., Ylinen, P., Konttinen, Y.T., and Tallroth, K. Foot and ankle fusions in Charcot-Marie-Tooth disease. Arch. Orthop. Trauma Surg. 112:175-179, 1993.
20. Scranton, P.E., Jr. Results of arthrodesis of the tarsus: talocalcaneal, midtarsal, and subtalar joints. Foot Ankle, 12:156-164, 1991.
21. Southwell, R.B., and Sherman, F.C. Triple arthrodesis: a long-term study with force plate analysis. Foot Ankle, 2:15-24, 1981.
22. Stephens, M., and Saleh, J. Calcaneal dome osteotomy: a new procedure for revising triple arthrodesis. Foot Ankle Int. 15:368-371, 1994.
23. Wetmore, R.S., and Drennan, J.C. Long-term results of triple arthrodesis in Charcot-Marie-Tooth disease. J. Bone Joint Surg. 71A:417-422, 1989.
24. Wukich, D.K., and Bowen, J.R. A long-term study of triple arthrodesis for correction of pes cavovarus in Charcot-Marie-Tooth disease. J. Pediatr. Orthop. 9:433-437, 1989
Fig. 1. Algorithm for surgical correction of failed triple arthrodesis.
Fig. 2. Correction of a varus hindfoot by closing wedge osteotomy through the talocalcaneal fusion mass (A) or by closing wedge osteotomy through the calcaneal tuberosity (B). Transverse tarsal osteotomy (C) is necessary when correcting varus by an osteotomy through the fusion mass (B).
Fig. 3. Axial view of closing wedge osteotomy with lateral translation of the calcaneal tuberosity for varus hindfoot (A) and axial (B) and lateral (C) views of screw placement across osteotomy through calcaneal tuberosity. [Reprinted by permission.]
Fig. 4. Fixed forefoot supination with location of transverse osteotomy through fusion mass, with derotation of forefoot (A) and fixation across osteotomy after forefoot derotation (B).
Fig. 5. Guide pin placement for closing wedge osteotomy through the talonavicular and calcaneocuboid fusion mass for excessive abduction of the forefoot (A) and fixation across osteotomy site after correction of abduction deformity (B).
Fig. 6. Plantar-based closing wedge osteotomy for correction of a rocker-bottom deformity (A) and fixation across osteotomy site after removal of plantar osteophyte and apposition of bone (B).
Fig. 7. AP ankle radiograph with lead-free solder marker positioned from the medial malleolus, around the heel, to the lateral malleolus to assess true hindfoot varus and valgus. The central axis of the marker is determined by marking two points midway between the medial and lateral arms of the tape, 2.5 and 5.5 cm from the most inferior portion of the tape (cross-hatches). The axis of the tibia is determined by marking two points in the central tibia 3 and 9 cm proximal to the plafond. In this example, the heel is in neutral (0o).
Fig. 8. Preoperative AP (A) and lateral (B) and postoperative AP (C) and lateral (D) radiographs for correction of an abduction deformity by medial closing wedge osteotomy. Note the preservation of the navicular and cuboid bone stock for screw placement.
Fig. 9. Preoperative AP (A) and lateral (B) and postoperative AP (C) and lateral (D) radiographs for correction of an adduction deformity by lateral closing wedge osteotomy. Note the substantial improvement in the talus-first metatarsal angle. A closing wedge osteotomy was also performed through the talocalcaneal fusion mass to correct hindfoot varus.
Fig. 10. Preoperative AP (A) and lateral (B) postoperative AP (C) and lateral (D) radiographs of a patient with severe rocker-bottom deformity. A plantar-based closing wedge osteotomy has resulted in excellent correction in the talus-first metatarsal angle.