Abstract
Background:
The results of first metatarsophalangeal (MTP) joint arthrodesis as a specific treatment for failed hallux valgus correction has not been previously reported. We evaluated the results of first metatarsophalangeal (MTP) joint arthrodesis as a treatment for failed hallux valgus surgery.
Methods:
The records of the senior author (MJC) were reviewed to identify patients treated for symptomatic failed hallux valgus procedures with arthrodesis of the first metatarsophalangeal joint over a 20-year period. Twenty-nine patients (33 feet) were available for followup examination. The patients completed a visual analog pain score, the American Orthopaedic Foot and Ankle Society (AOFAS) Hallux Metatarsophalangeal-Interphalangeal Scale, and a patient satisfaction scale at the final followup. Radiographs were obtained of both feet.
Results:
The average followup was 8 (1 to 22) years. The average pain score improved from 7 to 3 points. The mean AOFAS score was 73 points at followup. Patient satisfaction was excellent for 13 feet (39%), good in 11 (33%), fair in eight (24%), and poor in one (3%). The mean hallux valgus angle was 16 degrees, with an intermetatarsal angle of 8 degrees. There were three asymptomatic and one symptomatic nonunions. Twenty-two feet (67%) had corrective procedures performed on the lesser toes at the time of the first MTP joint arthrodesis.
Conclusions:
First MTP joint arthrodesis is a reliable option for revision after failure of surgical treatment of hallux valgus. This procedure can be used to treat a number of failure modes following initial hallux valgus surgery including recurrence, hallux varus, cock-up deformity, degenerative arthritis of the MTP joint, and associated transfer lesions beneath the lesser metatarsals. First MTP joint arthrodesis can be used after failed proximal and distal osteotomies, arthrodesis of the metatarsocuneiform joint, McBride procedure, exostectomy, and resection arthroplasty. Revision procedures have poorer clinical outcomes than successful primary hallux valgus procedures.
INTRODUCTION
Many surgical procedures have been used to correct hallux valgus deformities. 20 No one procedure is universally successful in all patients, and even a technically well-performed surgery may fail. 25 First ray surgery accounts for 66% of all failed forefoot procedures. 15 A failed hallux valgus procedure is a challenging problem because not only must the subsequent deformity be treated, but concomitant lesser toe and lesser metatarsophalangeal (MTP) joint deformities also must be corrected.
The use of a first MTP joint arthrodesis for failed hallux valgus procedures has been proposed by several authors. 4,6,9,16,19 –23,26,28,29 While results of primary arthrodesis for hallux rigidus or hallux valgus have been well documented, 5,7,11,13 long-term results of salvage of failed hallux valgus procedures have not been reported. Previous studies have included patients with multiple diagnoses, 4,6,19,22 describe surgical techniques without inclusion of a substantial number of patients, 3,28,29 reported salvage only after the Keller procedure, 9,18 or reported correction of only hallux varus. 26 We are unaware of any long-term outcome studies evaluating MTP arthrodesis for the treatment of failed hallux valgus procedures.
The purpose of the present study was to determine the results of arthrodesis of the MTP joint of the hallux as treatment for failed hallux valgus surgery by evaluation of the subjective and objective results of a single surgeon (MJC).
MATERIALS AND METHODS
The records of the senior author (MJC) were examined to identify patients who had an arthrodesis of the first MTP joint between 1983 and 2003. After institutional review board approval was obtained, the patients’ chart notes, radiographs, and operative reports were reviewed. By chart review, 42 patients were identified who had MTP joint arthrodesis after a primary bunion procedure. Seven patients had died since their surgery. Of the living 35 patients, three could not be located. Two patients had moved and were too distant to allow clinical evaluation. Both of these patients were contacted by telephone and expressed satisfaction with their procedures, but are not included in the study. One patient declined to participate in the study. The 29 remaining patients had undergone 25 unilateral and four bilateral procedures (33 feet). Two patients evaluated in the study for a unilateral procedure subsequently had the contralateral first ray corrected. The followup duration was not sufficient (less than 1 year) to include these patients in the bilateral group, and their results represent their status before the contralateral foot procedure.
The original procedures were identified in broad categories: exostectomy (three), McBride procedure (six), distal osteotomy (six), proximal osteotomy (11), resection metatarsophalangeal joint arthroplasty (five), arthrodesis of the MTP joint (one), and arthrodesis of the metatarsocuneiform joint (one). The time between the primary operative procedure and the salvage arthrodesis averaged 10.7 years (9 months to 35 years). Contributing medical conditions included diabetes (three), inflammatory arthritis (two), peripheral neuropathy (one), and pseudogout (one). Nine of the 33 (27%) had undergone at least one additional reconstructive surgery of the first ray before the salvage arthrodesis.
The average age at the time of salvage first MTP joint arthrodesis was 62 (29 to 89) years. Four of the 29 (14%) patients were men. The patients reported an average of 64 months of symptoms (2 to 240 months) before the salvage procedure. The average age at the time of followup was 69 (31 to 101) years.
The initial preoperative complaints included pain in 27 feet (82%), deformity in 21 feet (64%), shoewear difficulties in 14 (42%), and transfer metatarsalgia in eight feet (24%). Recurrence of hallux valgus was noted in 18 (55%) and hallux varus was noted in eight feet (24%). The average hallux valgus angle in patients with recurrence was 34 degrees. The average hallux varus angle in patients with hallux varus was 11 degrees. Other diagnoses included intractable plantar keratosis (11 feet, 33%), cock-up deformity (eight feet, 24%), degenerative arthritis (four feet, 12%), nonunion of arthrodesis (three feet, 9%), and metatarsus primus elevatus (one foot, 3%). Lesser toes complaints accompanied the hallux problems in 27 feet (82%).
An arthrodesis technique in which threaded Steinmann pin fixation that crossed the interphalangeal and the MTP joints was used in the five patients with the longest followup. 3 The remaining 28 patients were treated with conical reaming and internal plate fixation. 6 A dorsal plate either of Vitallium, stainless steel, or titanium provided internal fixation of the metatarsophalangeal joint. 12 Intercalary grafts were used in patients with previous resection arthroplasty of the MTP joint. Twenty-three of the 33 (70%) had additional corrective procedures at the time of the arthrodesis of the first MTP joint, 22 (96%) of which involved the lesser toes. These procedures included hammer toe corrections, Weil metatarsal osteotomies, flexor tendon transfers, Du Vries condylectomies, and resection arthroplasty of the lesser MTP joints.
The components of the American Orthopaedic Foot and Ankle Society (AOFAS) instrument, including pain, deformity, activity, shoewear problems, and range of motion, that could be found in the patient records were recorded. 27 A retrospective AOFAS score was not calculated. 27 The patient's level of pain, activity, length of symptoms, and complications before the arthrodesis were recorded. Two pain scores were used for evaluation. The visual analog pain score was used by most patients. The earliest cases predated the use of the visual analog pain scale in the senior surgeon's practice. For these early patients, pain was subjectively identified as none, mild, moderate, severe, and quite severe. These subjective descriptions were then assigned the values of 0, 3, 6, 9, and 10, respectively, for the purposes of this study. The average preoperative pain score was 7 for those patients with visual analog pain scales available and 7 for those patients using the old scale. Patients also were asked the distance they could routinely walk prior to and following the arthrodesis. Those patients who reported walking 1 mile were recorded as 12 blocks. Those patients who reported unlimited were listed as 36 blocks. Each patient was interviewed and examined by the same orthopaedic surgeon (JSG). The average followup was 8 (1 to 21) years.
During the followup evaluation, it was evident that a number of patients had multiple other complaints. Patients were encouraged to answer questions about pain based on the foot under consideration. When patients were unable to isolate their concerns to the involved foot, their answers were recorded verbatim. A number of patients had activity limitations that were not related to their foot at the time of followup. Activity levels were recorded as reported by the patient.
Patients were asked to rate their satisfaction with the result of the surgery according to a previously published scale.
4
In that scale, a result is considered
The physical examination evaluated function, alignment of toes, callus formation, gait, ligament laxity, and mobility of the first ray. For each patient, a Klaue device was used to measure the mobility of the first ray. The device is a modified ankle-foot-orthosis with an attached adjustable micrometer to measure the displacement of the metatarsal head in the sagittal plane. 14 The average of 10 trials with this device was recorded. An excess of 9 mm of mobility was considered abnormal. 17 The pronation of the first ray was evaluated and recorded using a 4-point scale in which none equals 1, mild equals 2, moderate equals 3, and severe equals 4. Harris mat studies were used to observe the pressure beneath the hallux and the lesser metatarsals.
The Beighton criteria 1 were used to measure ligamentous laxity (Table 1). This scale assigns values to specific physical findings.
Each patient's preoperative and followup radiographs were examined. Weight bearing anteroposterior and lateral radiographs were obtained on both feet of each patient at the time of followup. On the anteroposterior view the mid-diaphyseal line was drawn for the first and second metatarsals and the proximal phalanx. This line was defined by the midpoint between the two cortices for two locations on each bone. 10 These locations are defined as listed in Table 2. The distal point on the first metatarsal was moved proximal if the prior medial eminence resection extended more than 2 cm from the joint line. In patients with a previous Keller resection arthroplasty, the proximal phalangeal reference point was the most proximal point of the bone. The mid-diaphyseal lines were used to measure the intermetatarsal angle and the hallux valgus angle (Figure 1). The same reference points were used on the lateral images to define the inclination of the first metatarsal and the postoperative dorsiflexion angle of the arthrodesis. The plantar reference for the first metatarsal inclination was the plantar aspect of the sesamoids and the most plantar aspect of the calcaneus (Figure 2).
Ligamentous laxity as determined by the Beighton criteria 1
The points are added to arrive at one total. Patients can score between 0 and 9 points. Values over 5 points are considered to represent ligamentous laxity.
The mid-diaphyseal lines are drawn for the first and second metatarsal and the proximal phalanx. The hallux valgus and 1–2 intermetatarsal angles are indicated.
Measurement locations used to define the mid-diaphyseal line in centimeters from the adjacent joint
Radiographs also were used to examine the joints proximal and distal to the arthrodesis site. Fitzgerald's rating system was used to assess the interphalangeal joint as previously described. 5 The thickness of this joint's cartilage was measured using a digital caliper (L.S. Starrett Co., Athol, MA; measuring error 0.3 mm) at three points on both the AP and lateral images (Figure 3). The averages of these values were compared to the preoperative films and the contralateral foot films. The metatarsocuneiform joints were evaluated for degenerative changes. Each joint was assigned one of four categories: 1) no degenerative changes, 2) degenerative changes more than 0.5 mm narrowing, 3) joint narrowing of 0.5 to 1.0 mm, or 4) severe degenerative changes. 7
The plantar reference line is referenced off of the plantar aspect of the sesamoids and the most plantar aspect of the calcaneus. The mid-diaphyseal line is drawn for both the proximal phalanx and the metatarsal. The dorsiflexion and the metatarsal inclination angles are indicated.
Statistical Analysis
Comparative statistical analysis was performed using Excel© software 2003, SP1 (Microsoft Corp., Seattle, WA).
The students t-test was performed on continuous data.
RESULTS
The average visual analog pain score improved from 7 to 3 points. The mean AOFAS Hallux Metatarsophalangeal-Interphalangeal scale score was 73 (range 55 to 90) points. There was no difference between the average AOFAS score for those who had previous attempts to correct the failed bunion surgery and those without previous attempts at correction. Subjective patient satisfaction was reported as excellent in 13 feet (39%), good in 11 (33%), fair in eight (24%), and poor in one (3%). Patients were satisfied with pain relief in 26 feet (79%), correction of deformity in 29 (88%), tolerance of shoewear in 23 (70%), and cosmesis in 22 (66%). Twenty-three (79%) of the 29 patients stated they would have the procedure again.
Measurement locations on the interphalangeal joint of the first ray (arrows). The middle arrow on the AP radiograph is located where the mid-diaphyseal line of the proximal phalanx crosses the joint.
The average ambulatory capacity increased from 17 to 19 blocks. Only four (12%) patients had a decrease in ambulatory ability as measured by the maximal blocks walked. The average age at the time of followup of the patients with a decrease in blocks walked was 81 years (range 65 to 101). Ten patients (30%) showed an improvement in the number of blocks walked. The average age of the patients showing increased ambulatory ability was 75 (range 61 to 82) years.
All the feet left a clear imprint of the hallux on the Harris mat pressure study at final followup. Sixteen (48%) of the feet demonstrated increased pressure beneath the lesser metatarsals of which half were associated with metatarsalgia. One foot showed increased pressure on the lateral border of the foot. Twelve feet (36%) had calluses beneath the lesser metatarsals on physical examination. Nine of the feet with these calluses had increased pressure on the Harris mat studies.
According to the Beighton scale, 19 patients (66%) demonstrated no evidence (score of 0) of any laxity on the 9-point scale. Five patients had some positive findings on the Beighton scale (scores of 1 to 5). Five patients scored above 5 on the 9-point scale and were considered to have generalized ligamentous laxity. The five patients with ligamentous laxity had AOFAS scores of 73, 65, 65, 79, and 75 (mean 71).
Two patients (7%) had hypermobile (more than 9 mm displacement) first rays based on Klaue examinations. These two patients demonstrated significant laxity, scoring 7 on the Beighton scale. The patients scoring 7 or higher on the Beighton ligamentous laxity scale had increased Klaue measurements (10, 11, and 7 mm). No patient with a ligamentous laxity score of 6 or less had more than 6 mm of motion on the Klaue examination. The two patients with hypermobility had AOFAS scores of 73 and 65.
Mean Radiographic results
IP = interphalangeal.
The postoperative radiographic results (Table 3) demonstrated an average hallux valgus angle of 16, a first intermetatarsal angle of 8 degrees, and a dorsiflexion angle at the arthrodesis site of 22 degrees. Only one foot exhibited hallux varus (5 degrees), and only one foot had a hallux valgus angle of more than 30 (35) degrees. There were four (12%) radiographic nonunions, but only one was painful. The painful nonunion was treated with a successful revision arthrodesis.
Eight feet demonstrated a 1-point increase in the Fitzgerald arthritis score of the interphalangeal joint by 1 point and one foot had an increase of 2 points. When comparing the operated foot with the contralateral foot, there were as many nonoperated as operated feet with radiographic signs of degenerative changes. When threaded Steinmann pins were used as internal fixation, the Fitzgerald scores were higher, averaging 2.0 compared to 1.4 for all feet. The measured average thickness of the interphalangeal joint was 1.5 mm on the operated side and 1.4 mm on the nonoperated side. (p = 0.96). Only one foot had painful degenerative changes of the interphalangeal joint.
The metatarsocuneiform joint was noted to have degenerative changes in five feet. Two of these feet showed matching degeneration in the contralateral foot, and one of the feet was in a patient who had undergone bilateral procedures (one foot showed changes the other did not). The two remaining feet had grade 2 degenerative changes and were asymptomatic.
Pronation was absent in 28 feet (85%), mild in four feet (12%), moderate in one foot (3%), and severe in no feet. No patient had pain with stress of the first MTP joint arthrodesis site or identified the arthrodesis site as a source of pain.
Complications included four nonunions (12%) and two feet (6%) with ingrown nails. Ten (30%) feet required additional procedures related to the first metatarsal fusion. Five of these additional procedures were for removal of implants. The frequency of implant removal is related to the use of high-profile plates early in the study. More recent procedures with low-profile plates have not required implant removal (nine feet). Three feet had ingrown nails that required operative treatment, one foot had a revision of the fusion for a symptomatic nonunion, and another foot required an Akin procedure to correct the alignment of the first toe.
DISCUSSION
The failed procedures that comprise this study represent the referral practice of a single surgeon (MJC) and data should not be used to indicate the typical results of a particular procedure used for correction of hallux valgus. The original operative techniques are included only to illustrate the range of procedures that can be corrected with an arthrodesis of the first MTP joint. The small numbers of individual procedures precluded a determination of success of salvage of any one individual procedure. Failure of hallux valgus correction was not always treated with arthrodesis. Those without degenerative changes and mild hallux valgus or varus were treated with either soft-tissue procedures or proximal osteotomies combined with distal soft-tissue procedures.
The senior author (MJC) previously reported the results of arthrodesis as a primary procedure to treat hallux valgus deformities. 7 The patients were similar in age (69 years for the primary arthrodesis study and 62 in the current study). The average followup of 8 years was similar in both studies. The radiographic postoperative results also were similar, with mean hallux valgus angles of 20.8 degrees for primary arthrodesis and 16 degrees for revision surgery and mean 1–2 intermetatarsal angles of 11.8 degrees for primary arthrodesis and 8 degrees for revision surgery. The dorsiflexion angle of the arthrodesis of the MTP joint was 22 degrees in both studies. There were similar occurrences of nonunions and revision surgery in the two series. While the mean activity increased in both studies, four patients in the revision study decreased their activity level, while no patient in the primary study reported decreased activity.
The subjective scores demonstrated a clear difference between the primary and revision procedures (primary, 84% excellent, 16% good; revision, 39% excellent, 33% good, 24% fair, and 3% poor). The postoperative AOFAS scores also reflected the difference between the two groups: patients with a primary procedure scored 86 points, and those with a revision scored 73 points.
Machacek et al. 18 has reported a similar group of patients who had arthrodesis of the first MTP joint for failed Keller resection arthroplasty. Their patients were similar in age (64 years) and followup averaged 13 years. Their reported results for arthrodesis were similar (45% excellent, 34% good, 3% fair, and 17% poor) with a mean AOFAS score of 76.
Coetzee et al. 2 reported the use of the Lapidus procedure to treat failed hallux valgus. They reported an average AOFAS score of 88 at 24 months. The AOFAS score allows 10 points for first MTP joint motion. Thus, the maximal score our patients could achieve after MTP joint arthrodesis was 90 points. If these 10 points are excluded, the average score for patients treated with the Lapidus procedure for failed hallux valgus falls into the same range reported by Machacek et al. 18 and in the current study.
The AOFAS scores in these studies of revision procedures were lower than those reported for primary procedures. 2,7,16 This corresponds to our general impression from treating patients with failed hallux valgus procedures that patients have poorer clinical results and lower satisfaction levels after revision procedures than after primary procedures. Despite the lower AOFAS scores after revision compared to those after primary hallux valgus procedures, arthrodesis appears to be a reasonable method of treatment of a failed hallux valgus procedure. Most (79%) of our patients stated that they would have the procedure again. We believe this is indicative of the disabling nature of a failed hallux valgus procedure.
The goals of a first MTP joint fusion are pain relief and restoration of the load bearing capacity of the first ray. None of our patients complained of pain at the first MTP joint, and the Harris mat imprints confirmed that all patients were bearing weight through the first ray. However, almost half of the feet demonstrated increased pressure beneath the metatarsal heads of the lesser toes. Eight of the 16 patients with increased pressure beneath the metatarsal heads reported pain in this location. It appears that while the goal of restoring first ray load bearing was accomplished, subtle changes in the mechanics of the foot caused overload of the lesser metatarsals.
The alignment of the first ray showed consistent and durable results. The hallux valgus and 1–2 intermetatarsal angles and dorsiflexion of the arthrodesis were consistent with the results reported for primary first MTP joint arthrodesis. 7 Most (85%) of the first rays showed no pronation deformity. Minimal hallux varus occurred in one patient. The conical reaming used to prepare the joint results in curved coplanar surfaces that allow exacting control over the alignment of the arthrodesis. Pronation, dorsiflexion, and hallux valgus can all be adjusted without further altering the surfaces with repetitive bone cuts.
First Ray Mobility
First ray hypermobility has been reported to be a contributing factor in the failure of hallux valgus procedures. 2 Although it is clear that an axial plane deformity at the metatarsocuneiform joint develops with hallux valgus, instability in the sagittal plane has not been documented. Recent studies have not demonstrated hypermobility after the correction of hallux valgus deformity 7,8 and concluded that correction of the hallux valgus restores the stability of the first metatarsocuneiform joint and the first ray. Because of the retrospective nature of this study, preoperative mobility measurements could not be obtained. We did examine the patients at their followup examination to see if any component of hypermobility remained.
At the final followup, two patients (7%) had hypermobile (more than 9 mm of displacement) first rays based on examination with the Klaue device. Both patients reported excellent results with pain scores of 0 and 1 point on the visual analog pain scale and both had a successful fusion of the MTP joint. One patient had a low AOFAS score (65 points), but we attributed this to severe deformity of the contralateral foot and a concurrent debilitating illness. Both patients with hypermobility scored 7 of a possible 9 points on the Beighton ligamentous laxity test. Only one patient scored higher on the Beighton test (9 points), and this patient had displacement of 7 mm. The patients with ligamentous laxity had similar objective and subjective results as patients without signs of hypermobility. The small number of patients with hypermobility in our sample of failed hallux valgus procedures leads us to conclude that first ray hypermobility did not play a substantial role in failure of the hallux valgus procedures in this group of patients. While small numbers did not allow for statistical analysis, hypermobility did not change the clinical results of patients who underwent first MTP joint arthrodesis.
Degenerative Changes of the Interphalangeal and Metatarsocuneiform Joints
After first MTP joint arthrodesis, the first interphalangeal and the first metatarsocuneiform joints are at risk for degenerative changes. Only one patient in our study reported symptomatic changes in the interphalangeal joint, and no patient reported symptoms at the metatarsocuneiform joint at an average followup of 8 years. Followup radiographic examination demonstrated progression of degenerative changes at the interphalangeal joint in nine (27%) feet; however, just as many nonoperative feet demonstrated similar degenerative changes. Only two feet (6%) demonstrated degenerative radiographic changes at the metatarsocuneiform joint on only the operated side, and these were asymptomatic. We concluded that, degenerative change in adjacent joints was not an issue after first MTP joint arthrodesis and was more likely related to the advanced age of the patient population.
Limitations of the Study
This study is limited by its retrospective nature. Preoperative AOFAS scores would have improved the ability to compare the results with other studies. The transition from a subjective description to the visual analog pain scale made the comparison of pain relief difficult. The match between preoperative pain scores (average of 7 for both scales) indicates that our interpretation was reasonable.
In reporting the subjective aspects of the clinical scores, it often was difficult for our patients to exclude complaints not related to the arthrodesis procedure. In general, patients could not separate cosmetic, functional, or pain complaints related to lesser toes from those related to the first ray. The failed hallux valgus procedure was believed to have affected the lesser toes in a large number of patients (82%) and most (96%) had additional procedures at the time of the salvage procedure. Although it is reasonable to include complaints related to the lesser metatarsals in the evaluation of the procedure, in some patients unrelated medical conditions affected their final subjective report. The pain score in particular was affected by conditions such as peripheral neuralgia and sciatica.
The clinical instruments available to evaluate patients undergoing foot and ankle procedures are nonspecific. The clinical results contain confounding elements. Saltzman 24 reported steady declines in the AOFAS scores after total ankle replacement unrelated to the performance of the ankle. Activity and pain scores were affected in this study by patient comorbidities. A number of our patients were unable to distinguish the pain in their feet from other sources of discomfort. Activity levels were reduced in a number of patients because of spine conditions, other unrelated lower extremity conditions, or deconditioning associated with advanced age. Given the long followup period of this study, up to 21 years, it is not surprising that there was a general decline in health in the patients. The average patient was 62 years old at the time of the salvage procedure and 69 years old at the time of the followup examination. Because of the small number of patients, we are unable to account for the contribution of the comorbidities to the clinical scores. While the AOFAS scores were recorded exactly as provided by the patient, the relatively low scores do not seem to match the higher patient satisfaction with the procedure.