We retrospectively reviewed the records of forty-one patients who had been seen at The Union Memorial Hospital in Baltimore, Maryland, between 1986 and 1992, because of pain secondary to osteoarthrosis and deformity after an injury of the tarsometatarsal joints. Of these forty-one patients, one had been managed initially at our institution and forty had been referred for evaluation after initial management at another medical facility. The diagnosis of injury of the tarsometatarsal joints was made by means of a review of previous medical records, the patient's history, and the results of physical and radiographic examinations.
Regardless of the time since the injury, we initially treated the pain with various non-operative modalities, including non-steroidal anti-inflammatory medication, an accommodative or functional orthosis, modifications of the shoe, a polypropylene ankle-foot orthosis, and physical therapy. An arthrodesis was recommended if symptoms persisted after three to six months of non-operative care. Six patients had subjective improvement after a variety of non-operative interventions. Three patients had no improvement but chose not to have an operation. In seven of these nine patients, the osteoarthrosis involved the second and third metatarsocuneiform joints; in the other two, it involved the first metatarsocuneiform joint. The injuries and the degree of pain in these nine patients were similar to those in the remaining thirty-two patients, who were managed with an arthrodesis for persistent symptoms. These thirty-two patients (seventeen women and fifteen men), who were twenty to sixty-seven years old (mean, forty-four years old) at the time of the arthrodesis, formed our study group.
Initial Injury and Treatment
The initial injury was sustained six to 108 months (mean, thirty-five months) before the arthrodesis. The mechanism of injury was a fall for thirteen patients, a motor-vehicle accident for ten, a crush injury for five, and a minor twisting accident for four (Table I). Four patients (Cases 2, 6, 13, and 15) had multiple associated injuries, including fractures of the femur, the tibia, the ankle, the hip, or the spine. Seven patients (Cases 2, 5, 6, 10, 13, 19, and 24) had associated injuries of the cuneiforms, the cuboid, and the metatarsals.
Nine patients had had operative treatment of the initial injury. Four of these patients (Cases 14, 19, 29, and 32) had had closed reduction and percutaneous fixation with Kirschner wires; in three of the four (Cases 19, 29, and 32), the fracture or dislocation was malreduced—that is, there was still more than 15 degrees of abduction of the first metatarsal or more than two millimeters of lateral shift of at least one metatarsal, or both, after the fixation15. The other five patients who had been managed operatively had had open reduction and internal fixation: four (Cases 2, 7, 15, and 28) with Kirschner wires and one (Case 1, the only patient to have had the initial injury treated at our institution) with screws. In three patients (Cases 2, 7, and 28), there was still more than two millimeters of displacement of the metatarsals after fixation. One of these patients (Case 28) had comminuted fractures of the metatarsals and cuneiforms that had been secured with Kirschner wires through a large open wound.
Twenty-three patients had had non-operative treatment of the initial fracture or dislocation. We examined the radiographs of the initial injury for seventeen of these twenty-three patients. The injury of the tarsometatarsal joints had been either misdiagnosed or missed completely for ten of these seventeen patients: five patients (Cases 8, 20, 23, 30, and 31) had been managed for a sprained foot, and the other five (Cases 5, 10, 16, 22, and 25) had been managed for a fracture of the metatarsals, the cuneiforms, or the cuboid. The dislocation of the tarsometatarsal joints had not been identified in any of these ten patients.
Eleven of the twenty-three patients who had been managed non-operatively had had immobilization in a cast with no attempt at reduction of the displaced fracture or the dislocation; seven of these eleven patients had been allowed to bear weight as tolerated while wearing the cast, and the other four had been instructed to remain on crutches for a mean of three weeks. The remaining twelve patients had been managed with immobilization in a cast after attempted reduction of the fracture or the dislocation. Ten of these twelve patients had some degree of malalignment (two to six millimeters of displacement of the metatarsal or lateral angulation of more than 15 degrees, or both), and two patients had no discernible radiographic malalignment.
Evaluation
The decision regarding the preferred extent of the arthrodesis was made on the basis of the location of the pain and the radiographic appearance of the joints. We divided the tarsometatarsal articulation into three columns: the medial column (the first metatarsocuneiform joint), the middle column (the second and third metatarsocuneiform joints), and the lateral column (the fourth and fifth metatarsocuboid joints). We used passive manipulation of the midfoot, which involves simultaneous pronation and abduction of the forefoot (Fig. 1), to determine the location of maximum pain. We did not inject a local anesthetic to determine which joints were painful, as these joints are small and selective anesthesia of at least one joint was not thought to be sufficiently accurate. Although technetium bone scans may be useful for localization of the involved joints, they were not found to be clinically useful in the present study, as increased uptake was often noted in joints that were not painful. This was particularly true with respect to the lateral column, in which arthrodesis was rarely performed but increased uptake was often noted. It has been our experience that these additional diagnostic modalities are unnecessary in the determination of which joints are painful.
Preoperative and postoperative anteroposterior and lateral radiographs of both feet were made with the patient bearing weight. The talus-first metatarsal angle and the distance from the medial cuneiform to the floor as measured on the radiographs of each injured foot were compared with those on the radiographs of the contralateral, uninjured foot.
Operative Technique
Two methods of arthrodesis were used. An in situ arthrodesis, with no attempt at realignment, was performed for eight patients who had slight deformity. Twenty-four patients had an arthrodesis after an attempted realignment of the forefoot in both the sagittal and transverse planes.
The arthrodesis was performed with general anesthesia for twenty-two patients and with a local ankle block for ten. Twenty-four patients were managed with a bone graft, from the ipsilateral calcaneus (thirteen patients) or the ipsilateral iliac crest (eleven patients). Bone-grafting was necessary when the joints could not be aligned with osseous contact. The type of graft was selected on the basis of the size of the defect created by débridement of the joints. Intraoperative radiographs were made when any realignment of the forefoot was performed, and fluoroscopy was used routinely to determine the position of the forefoot after the application of the internal fixation device.
The arthrodesis was performed with screws for all patients; however, the type of screw varied. Initially, 4.0-millimeter cancellous-bone and 3.5-millimeter cortical-bone screws were used. The decision to use cortical-bone rather than cancellous-bone screws was based on the size of the metatarsals. More recently, 3.5-millimeter cortical-bone screws and 4.5-millimeter cannulated cancellous-bone screws were used because of the ease and accuracy of placement. If the joints of the lateral column were in a position of malreduction, they were mobilized by resection of scar tissue, were realigned, and were secured with either Kirschner wires or screws. This fixation was temporary and was removed once weight-bearing had begun, approximately eight weeks after the operation. Only two patients had an arthrodesis of the lateral column. Five patients (Cases 18, 26, 27, 30, and 31) had an arthrodesis of the medial column alone; five (Cases 5, 11, 16, 20, and 21), of the middle column alone; twenty, of the medial and middle columns; and two (Cases 3 and 19), of all three columns. Although the arthrodesis of the middle column included only the second metatarsal-middle cuneiform joint in some patients, for practical purposes we applied this designation to both the second and third metatarsals. An isolated arthrodesis of the third metatarsocuneiform joint was not performed. In four patients (Cases 2, 10, 24, and 31), the arthrodesis also included the naviculocuneiform joint.
Nine patients had at least one concomitant procedure. Three patients (Cases 2, 14, and 26) had reconstruction of the posterior tibial tendon, one (Case 13) had an arthrodesis of the ankle, eight (Cases 3, 5, 6, 13, 14, 21, 26, and 28) had a claw-toe procedure, three (Cases 3, 5, and 13) had an excision of an interdigital neuroma, and one (Case 6) had an arthrodesis of the calcaneocuboid joint.
In Situ Arthrodesis
An in situ arthrodesis is indicated for patients who have slight deformity and osteoarthrosis that is limited to the medial or middle column, or both (Figs. 2-A, 2-B, 2-C through 2-D). Local anesthesia is often sufficient; if bone-grafting is necessary, local bone graft is preferred. One incision is made centered over the second tarsometatarsal joint. If only the medial column is to be involved in the arthrodesis, the incision is made dorsal to and centered over the shaft of the first metatarsal. The dorsocentral incision is made with care being taken to identify and protect the superficial and deep peroneal nerves, the dorsalis pedis artery, and the vertical descending arterial branch in the first web space. The joints to be included in the arthrodesis are exposed, the capsules are opened, and all fibrous tissue blocking reduction is removed. Digital traction and plantar flexion of the forefoot allows distraction of the joints. The cartilage and the subchondral bone are then removed with a beveled chisel. It is important to remove the plantar aspect of the joint in order to prevent dorsal angulation and malunion. The reduction is held temporarily with guide-pins or Kirschner wires, and radiographs are made to confirm the anatomical reduction. A high-speed burr is then used to debride the edges of the joint; any gap that is created is filled with local bone graft, which sometimes is obtained from the calcaneus. Additional bone graft is usually not necessary with in situ procedures. The size of the screw is selected according to the size of the metatarsal. We use 4.5 and 3.5-millimeter screws for the medial and middle columns, respectively. To avoid splitting of the dorsal cortex of the metatarsal, the hole must be beveled so that the screw can be countersunk carefully or the screw must be inserted in a proximal-dorsal to distal-plantar direction.
Arthrodesis with Realignment
For patients who have residual displacement and deformity of the forefoot, we believe that realignment of the forefoot is preferable to an in situ arthrodesis18. Although we know of no previous reports that have discussed the amount of displacement that necessitates realignment and arthrodesis, the morbidity after injury of the tarsometatarsal joints has been demonstrated to worsen with more than two millimeters of displacement or more than 15 degrees of malalignment, or both15. In the present study, any patient who had at least three millimeters of displacement or at least 15 degrees of malalignment in the transverse or sagittal plane was managed with realignment and arthrodesis (Figs. 3-A, 3-B, 3-C through 3-D). Typically, a flatfoot deformity is present in association with abduction of the forefoot and lateral translation and dorsiflexion of the metatarsals. The medial soft tissues may be attenuated and, occasionally, the lateral soft tissues may be contracted. Our goal in reducing the deformity was to restore alignment of the medial aspect of the base of the first metatarsal with the medial edge of the first cuneiform, to restore alignment of the medial aspect of the base of the second metatarsal with the medial edge of the second cuneiform in the transverse plane, and to align the long axis of the talus with the long axis of the first metatarsal in both the sagittal and transverse planes3,18 (Fig. 4). The incisions are planned according to the deformity. For a severe deformity, three incisions are made: one dorsomedial to the first metatarsal; one between the second and third metatarsals; and one dorsal to the fifth metatarsal, occasionally extending proximally to the calcaneus.
In patients who have severe abduction of the forefoot, the peroneus brevis tendon and other lateral soft tissues are contracted and must be lengthened. If the contracture is severe, we use a small external-fixation device intraoperatively to assist with the reduction, as described by Sangeorzan et al.17,18. A pin is inserted laterally into the fifth metatarsal and a second pin, into the calcaneus. If the pin that is inserted into the fifth metatarsal does not obtain secure fixation, it is advanced and inserted into the fourth metatarsal. The pins are inserted so as to converge slightly as an aid in the realignment. Before distraction, all of the joints to be included in the arthrodesis are mobilized by resection of the scar tissue, debris, and cartilage. A beveled chisel is used to denude the surfaces of the joint, but no bone is resected.
We find that the realignment is facilitated by initial correction of the position of the first metatarsal, which is performed by grasping the great toe with the hand and forcing it into varus while the base of the first metatarsal is pushed laterally with the thumb (Fig. 4). The first metatarsal is temporarily secured with a cannulated guide-pin that is inserted from the dorsal surface of the metatarsal proximally into the medial cuneiform. A large bone-reduction clamp is placed obliquely to close the gap between the base of the second metatarsal and the medial cuneiform. Both the second and the third metatarsal are secured with Kirschner wires that are placed through these bones and then into the respective cuneiforms, and a cannulated guide-pin is inserted from the medial cuneiform distally toward the base of the second metatarsal. Anteroposterior and lateral radiographs are made to confirm the corrected alignment.
The arthrodesis is performed with 3.5-millimeter cortical-bone screws introduced in a lag fashion or with partially threaded, cannulated 4.5-millimeter screws. It has been our experience that, despite substantial arthrofibrosis and deformity, there usually is no pain in the joints between the fourth and fifth metatarsals and the tarsal bones postoperatively, provided that the alignment has been corrected18. If the fourth and fifth metatarsals are displaced more than two millimeters, they are anatomically reduced and this position is maintained with smooth Kirschner wires4,5,12,20, which are removed with the resumption of weight-bearing at approximately eight weeks. The wounds are closed with 4-0 subcuticular sutures, and 5-0 chromic simple interrupted sutures are used for the skin.
A bulky compression dressing is applied and is worn for two weeks, and then a cast is applied. Weight-bearing is begun in a below-the-knee weight-bearing cast or a commercially available weight-bearing boot at approximately eight weeks; the cast or boot may be used for an additional four to eight weeks, depending on the healing of the arthrodesis as determined on the radiographs.
Analysis of Results
The results of the arthrodesis were evaluated objectively with clinical and radiographic examination and by comparison of the preoperative and postoperative scores on the 100-point clinical-rating scale developed by the American Orthopaedic Foot and Ankle Society for evaluation of the midfoot11. The preoperative scores were obtained retrospectively with use of a questionnaire and through review of the medical records.
Anteroposterior and lateral radiographs, made with the patient bearing weight, were used to evaluate the alignment and the presence of fusion. Measurements of the talus-first metatarsal angle in both the lateral and the anteroposterior plane and the distance from the medial cuneiform to the floor were used to assess alignment.
The statistical tests that were used to evaluate the data included the Student t test, a one-way analysis of variance with a Scheffé comparison, and Pearson correlations. The statistical software program that was used was SPSS/PC (SPSS, Chicago, Illinois).
The thirty-two patients were examined at a mean of fifty months (range, twenty-four to 105 months) after the arthrodesis, and thirty-one had fusion at the site of the arthrodesis. One patient (Case 14), who had been managed with an arthrodesis of the medial and middle columns, had fusion of the middle column but had a painless non-union of the first metatarsocuneiform joint. In addition to the arthrodesis of the tarsometatarsal joints, a reconstruction of the flexor tendon had been done concomitantly because of weakness of the posterior tibial muscle and an acquired flatfoot deformity. Although this patient had intermittent swelling and aching after the operation, she had no pain in the midfoot after nine months.
It was not possible to determine accurately when fusion occurred, as all of the patients were examined at four-week intervals. It was our impression, however, that fusion, as determined by the absence of warmth and swelling and by the presence of radiographic signs of trabeculation, occurred by eight weeks in sixteen patients, by ten weeks in ten patients, and by twelve weeks in five patients.
American Orthopaedic Foot and Ankle Society Score
The postoperative scores according to the 100-point clinical-rating scale developed by the American Orthopaedic Foot and Ankle Society for evaluation of the midfoot11 were calculated and then were compared with the preoperative scores, which had been determined retrospectively. Over-all, there was a significant difference between the mean preoperative score of 44 points (range, 22 to 64 points) and the mean postoperative score of 78 points (range, 47 to 98 points) (p = 0.02, Student t test). With the numbers available, we were unable to demonstrate a significant difference, with respect to the outcome, between the mean preoperative and postoperative scores according to the mechanism of injury or to the extent and location of the arthrodesis (one-way analysis of variance) (Table II).
The injury had been work-related in six patients (five of whom had a crush injury and one of whom fell). The mean preoperative and postoperative scores for these six patients were 41 points (range, 26 to 58 points) and 71 points (range, 51 to 88 points), respectively. The mean preoperative and postoperative scores for the twenty-six patients who had sustained a non-work-related injury were 44 points (range, 22 to 64 points) and 80 points (range, 47 to 98 points), respectively. With the numbers available, there was no significant difference between these two groups (p = 0.81, Student t test).
With the numbers available, we could not demonstrate a significant difference between the postoperative scores of the nine patients who had had at least one additional procedure at the time of the arthrodesis and those of the remaining twenty-three patients (p = 0.09, Student t test) (Table II).
The mean preoperative score for the five patients who had a revision procedure after the arthrodesis was 41 points (range, 26 to 52 points) and the mean score at the latest follow-up was 67 points (range, 47 to 83 points). Three patients had neuritis or a neural injury; the mean score was 43 points (range, 27, 50, and 52 points) preoperatively and 60 points (range, 54, 59, and 66 points) postoperatively. The preoperative and postoperative scores for the patient in whom a non-union of the medial column developed were 58 and 77 points, respectively. We could not demonstrate a significant difference between the postoperative scores of the five patients who had a revision and those of the twenty-seven patients who did not (p = 0.14, Student t test).
There was no correlation between the age of the patient at the time of the arthrodesis and the outcome (r = 0.03; p < 0.05, Pearson correlation).
Radiographic Analysis
Radiographic analysis was performed only for the twenty-four patients who had had realignment, reduction, and arthrodesis. Abduction of the metatarsals, or lateral angulation, as measured on the anteroposterior radiographs and expressed as the talus-first metatarsal angle (normal angle, approximately 0 degrees), improved from a mean of 13 degrees (range, 0 to 28 degrees) of abduction preoperatively to a mean of 4 degrees (range, 0 to 8 degrees) of adduction postoperatively. The height of the first metatarsal, as measured on the lateral radiographs and expressed as the lateral talus-first metatarsal angle (normal angle, approximately 0 degrees), improved from a mean of 16 degrees (range, 0 to 24 degrees) of dorsiflexion preoperatively to a mean of 6 degrees (range, 0 to 10 degrees) of dorsiflexion postoperatively. The distance from the medial cuneiform to the floor improved from a mean of twenty-four millimeters (range, seventeen to forty-two millimeters) preoperatively to a mean of thirty-six millimeters (range, twenty-four to forty-two millimeters) postoperatively. The mean postoperative distance from the medial cuneiform to the floor for the injured foot compared favorably with that for the contralateral, uninjured foot (thirty-nine millimeters; range, thirty-four to forty-five millimeters).
Complications
Neuritis or a neuroma of the deep peroneal nerve developed postoperatively in three patients (Cases 5, 7, and 28). One of these patients (Case 7) had been managed initially with open reduction and internal fixation with percutaneous Kirschner wires for a crush injury. Malunion and severe neuritis of the deep peroneal nerve developed. The neuritis subsequently was treated operatively before the patient was evaluated and managed by us. Although the neuroma of the deep peroneal nerve was identified and resected at the time of the arthrodesis, a repeat procedure with resection of the deep peroneal nerve and proximal transposition was necessary. A neuroma of the deep peroneal nerve in another patient (Case 5) led to a subsequent resection. Both of these patients continued to have hyperesthesias and dysesthesias postoperatively. The third patient (Case 28) had neuritis of the deep peroneal nerve; however, the painful symptoms resolved after conservative treatment, which consisted of the administration of twenty-five milligrams of amitriptyline at night for one month.
Metatarsalgia developed in two patients (Cases 3 and 19) as a result of malunion of at least one metatarsal. In one patient (Case 3), the metatarsalgia was thought to be due to malunion of the second metatarsal, which was plantar flexed, and it was treated successfully with a dorsal-wedge osteotomy of the second metatarsal. The second patient (Case 19) had an arthrodesis of all three columns (Fig. 5-A, 5-B, 5-C through 5-D). Pain developed postoperatively under the second and third metatarsals and was treated with a dorsal-wedge osteotomy of both of these metatarsals. The patient then had pain under the heads of the fourth and fifth metatarsals, which was in turn treated with dorsal-wedge osteotomies. The multiple osteotomies were successful in alleviating the discomfort, and the patient was able to return to construction work with minimum pain.
A superficial wound slough developed in one patient (Case 28), who had sustained a crush injury that was associated with soft-tissue loss and severe comminution of the fracture. The wound slough was treated successfully with a split-thickness skin graft in the immediate postoperative period. A superficial wound infection developed in another patient (Case 5) and was treated effectively with oral administration of 500 milligrams of cephalexin four times a day for ten days. There were no deep infections. One patient (Case 7) who had severe recurrent postoperative neuritis of the deep peroneal nerve also had signs and symptoms of reflex sympathetic dystrophy. Although the sympathetically mediated pain resolved with lumbar sympathetic blocks, physical therapy, and the administration of twenty-five milligrams of amitriptyline at night for two months, the symptoms of neuritis continued.
Five patients (Cases 3, 5, 6, 7, and 19) had at least one subsequent procedure: two (Cases 3 and 19) had a dorsal-wedge osteotomy for metatarsalgia, two (Cases 5 and 7) had resection of a neuroma, and two (Cases 5 and 6) had an additional arthrodesis. The additional arthrodeses were of the first metatarsocuneiform joint (Case 5) and of the fourth and fifth metatarsocuboid joints (Case 6) and were performed because of painful osteoarthrosis that had not been diagnosed previously.
A high prevalence of post-traumatic osteoarthrosis after tarsometatarsal fractures and dislocations, with or without adequate initial treatment, has been reported1,5,7,9,12,20. When non-operative treatment fails to relieve the symptoms of osteoarthrosis to an acceptable degree, arthrodesis of the painful tarsometatarsal joints is the treatment of choice10,18,19.
Johnson and Johnson reported on a series of fifteen patients for whom arthrodesis of the tarsometatarsal joints was performed with use of a dowel-graft technique to treat post-traumatic osteoarthrosis. There was no attempt to reduce the deformity, and the arthrodeses were performed in situ. Those authors reported two painful non-unions that necessitated repeat arthrodesis with inlay bone graft. They also reported a fair or poor result in five of their fifteen patients, as determined by the level of activity, pain, and radiographic union. Although the authors suggested that a decline in the quality of the result after the arthrodesis was associated with increasing age, the study group was too small to demonstrate significance10. The dowel-graft technique is not our preferred method of arthrodesis because the removal of a dowel creates additional instability of the metatarsocuneiform articulation and because this technique is not associated with an acceptable rate of success, as originally reported10.
Sangeorzan et al.18 reviewed the results for sixteen patients in whom osteoarthrosis developed after an injury of the tarsometatarsal joints and who had an arthrodesis performed with rigid internal fixation. Those authors identified an association between the outcome and the time between the injury and the arthrodesis, the alignment of the arthrodesis, the quality of the reduction, and whether the injury was work-related. With use of stepwise multiple-regression analysis, however, Sangeorzan et al. determined that the only useful factor for predicting outcome was the alignment of the arthrodesis18. Those authors recommended that the lateral column not be involved in the arthrodesis, and their findings indicate that an arthrodesis of the lateral column is not necessary for a good result.
The experience of Sangeorzan et al.18 was similar to ours in that many of our patients had radiographic evidence of osteoarthrosis of the fourth and fifth metatarsocuboid joints but had no pain in this location preoperatively or postoperatively. The reason for this is not well understood. Various authors have quantified the motion of the lateral column of the midfoot13,16. Interestingly, we found that the lateral column, which has the greatest motion (mean, 10 degrees) in the sagittal plane16, was the least painful. However, the second metatarsocuneiform articulation, which has the least motion (mean, 0.6 degree)16, was the most painful.
Only two patients in our study needed an arthrodesis of the lateral column. Interestingly, both of these patients (Cases 3 and 19), who had an extended arthrodesis of all three columns, needed a subsequent revision procedure with metatarsal osteotomies to treat metatarsalgia. A third patient (Case 6) had sustained an injury in association with abduction of the forefoot and compression of the metatarsocuboid and calcaneocuboid joints. The metatarsocuboid joints were not painful, and realignment of the forefoot was achieved with an interposition bone-block arthrodesis of the calcaneocuboid joint. Two years later, the patient began to have intractable lateral pain and had an arthrodesis of the metatarsocuboid joints. There was minimum lateral motion before the second arthrodesis, and we concluded that the pain was in some way aggravated by the decreased motion in the lateral column. Although these numbers are small, it may be inferred that the motion of the lateral column is important for optimum function. Therefore, whenever possible, we recommend that the lateral column not be included in the arthrodesis, particularly as most of these joints are asymptomatic despite radiographic evidence of osteoarthrosis.
This leads to the question of which columns should be included in the arthrodesis. We used clinical examination and plain radiographs to determine which joints were to be included. When a bone scan was made, there was diffuse uptake throughout the midfoot, often in locations that were painless. These scans are very sensitive to osseous turnover or increased blood flow secondary to inflammation from osteoarthrosis, but these findings do not always coincide with the painful joints. Although bone-scanning has been used by others19, we do not recommend its use or the use of other imaging modalities to determine the extent of the arthrodesis19.
Radiographic analysis demonstrated satisfactory postoperative alignment as compared with that of the contralateral, uninjured foot. Because our study did not include feet in which an in situ arthrodesis was performed in the presence of obvious malalignment, we cannot compare the two techniques. However, we believe that realignment of the midfoot contributes to the over-all good result, as demonstrated by the scores for the evaluation of the midfoot11.
We believe that when an arthrodesis of the tarsometatarsal joints is performed to treat symptomatic osteoarthrosis following a traumatic injury, it should involve realignment of the deformity of the midfoot (when such a deformity is present), the use of rigid fixation, and bone-grafting as needed. The patients in this study had marked improvement with respect to both pain and function after the arthrodesis, as determined with the clinical rating scale developed by the American Orthopaedic Foot and Ankle Society for the evaluation of the midfoot11.