- Email: [email protected]
The Effect of Hindfoot Realignment in Triple Arthrodesis
ORIGINAL RESEARCH
The Effect of Hindfoot Realignment in Triple Arthrodesis Brandon J. Child, DPM,1 Joel Hix, DPM,1 Alan R. Catanzariti, DPM, FACFAS,2 Robert W. Mendicino, DPM, FACFAS,3 and Karl Saltrick, DPM, FACFAS4 Triple arthrodesis is a useful form of surgical intervention for repair of advanced hindfoot deformity. Crucial to the success of triple arthrodesis is achievement of a suitable alignment of the hindfoot in relation to the leg and forefoot. A number of radiographic measurements can be used to describe the resultant alignment of the foot and, in this article, we present the results of a review of a series of 28 triple arthrodeses, in 24 patients, followed for a median duration of 19 (range 12 to 38) months. The results of this investigation showed clinically and statistically significant improvements in the anteroposterior talocalcaneal angle and talo-first metatarsal angles, and the lateral talo-first metatarsal angle; and triple arthrodesis realigned the hindfoot primarily by means of transverse plane movement of the talus over the calcaneus, and sagittal plane realignment occurred primarily by means of reduction of the downward pitch of the talus rather than by means of elevation of the pitch of the calcaneus. Only 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) symptomatic nonunion was observed. Correlation analyses showed that a normal (35 to 50 ) preoperative lateral talocalcaneal angle was moderately inversely correlated, and a normal (# 21 ) preoperative lateral talar declination angle was strongly positively correlated, with none to mild postoperative pain. Based on these findings, we concluded that triple arthrodesis satisfactorily realigned the hindfoot and reduced pain in patients with advanced arthrosis. Level of Clinical Evidence: 4 (The Journal of Foot & Ankle Surgery 48(3):285–293, 2009) Key Words: arthrosis, cavus, deformity, fixation, pes valgus, varus
The primary goals of triple arthrodesis are to provide stability, decrease pain, and reduce deformity; and appropriate realignment of the hindfoot is a key factor in achieving an acceptable surgical outcome. A well-aligned triple arthrodesis should ultimately result in a plantigrade foot that will support the ankle and reduce the risk of complications such as weight-bearing instability, the progression of Address correspondence to: Robert W. Mendicino, DPM, FACFAS, The Foot and Ankle Institute of Western Pennsylvania, 4800 Friendship Avenue, North Tower, First Floor, Pittsburgh, PA 15224. E-mail: rwmendicino@ faiwp.com 1 Third-Year Resident, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. 2 Director of Residency Training, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. 3 Chief, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. 4 Assistant Director of Residency Training, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. Financial Disclosure: None reported. Conflict of Interest: None reported. Copyright Ó 2009 by the American College of Foot and Ankle Surgeons 1067-2516/09/4803-0002$36.00/0 doi:10.1053/j.jfas.2009.02.006
foot and/or ankle deformity, and the development of secondary degeneration of adjacent joints. Although triple arthrodesis is generally known to result in favorable longterm results relative to pain relief and patient satisfaction (1–5), we feel that there has been limited attention paid to the importance of precise hindfoot realignment as it relates to the outcome of the procedure. Furthermore, poor outcomes have been reported in association with tarsal misalignment, such as fixed hindfoot varus (6–8). Although it is widely accepted that surgeons should avoid fusion of the hindfoot in varus, a thorough review of the biomedical databases revealed just a few reports (5, 9, 10) that quantitatively described the association of tarsal realignment with the outcome of a triple arthrodesis. Sangeorzan and colleagues (5) observed 44 triple arthrodeses in 40 patients, with an average of 4.9 years of follow-up, and noted an average improvement of 17 in the lateral talometatarsal angle, 11 in the lateral talocalcaneal angle, and 18 in the anteroposterior (AP) talometatarsal angle. Bednarz et al (9) observed 63 triple arthrodeses in 57 patients, at an average of 30 months follow-up, and noted an average improvement of 12 in the VOLUME 48, NUMBER 3, MAY/JUNE 2009
285
lateral talometatarsal angle, 7 in the lateral talocalcaneal angle, and 10 in the AP talometatarsal angle (P < .0001). Finally, Wenz et al (10) observed 6 patients with paralytic pes calcaneus at an average follow-up of 32 months following triple arthrodesis, and noted statistically significant improvements in the talocalcaneal angle, the calcaneal pitch, and the talo-first metatarsal angle. The purpose of the current investigation was to review subjective and objective outcomes in 24 patients who underwent 28 triple arthrodeses, with particular emphasis paid to the influence that hindfoot alignment had on the long-term measurement of radiographic angles and reduction of pain following the operation. Patients and Methods Chart Review The authors reviewed the records of 24 patients who had undergone triple arthrodesis (28 procedures) at the Foot and Ankle Institute of Western Pennsylvania, at The Western Pennsylvania Hospital, in Pittsburgh, Pennsylvania, during the period extending from September 2000 to May 2003. After obtaining approval from the Institutional Review Board, patient information was abstracted from the patient records and radiographic measurements were made using the preoperative and postoperative radiographic images, and a long-term follow-up pain questionnaire was dispensed and returned via the postal service. The questionnaire was sent to all patients who underwent triple arthrodesis during the above captioned time period. Radiographs were measured at a median duration of 19 months. Radiographic Assessment Preoperative and postoperative radiographs were compared to determine the degree of structural realignment of the foot. To be included in the case series, patients had to have at least 1-year follow-up notes and radiographs in the medical record. To assess forefoot-to-hindfoot relationships, AP talo-first metatarsal and talocalcaneal angles were used to determine the transverse plane alignment (Figures 1 and 2). On the lateral foot radiograph, the talo-first metatarsal, talocalcaneal, calcaneal inclination, and talar declination angles were measured (Figures 3–5). Furthermore, to correlate radiographic alignment with pain outcomes, the radiographic measurements were categorized relative to normal values (11) as follows: AP talocalcaneal angle 25 or less, AP talofirst metatarsal angle less than 15 , calcaneal inclination angle equals 15 to 21 , lateral talocalcaneal angle equals 35 to 50 , lateral talo-first metatarsal angle equals 5 to 5 , and talar declination angle 21 or less. Still further, the time to successful osseous union was also determined radiographically by assessing serial postoperative radiographs, and successful union 286
THE JOURNAL OF FOOT & ANKLE SURGERY
was considered present when there was evidence of trabecular bone crossing the arthrodesis sites. Pain Assessment The short-form McGill visual analog pain scale questionnaire (SF-MPQ) (12) was used to evaluate preoperative and postoperative pain. The SF-MPQ has 15 pain descriptors that are subjectively rated by the patient, and uses an intensity scale of 0 (none) to 3 (severe), with the minimum score being 15 (no pain) and the maximum score being 45 (severe pain). The SF-MPQ has been found to be reliable and to provide valid information related to pain in acute and chronic situations (12). Preoperative and postoperative pain was also categorized as none, mild, moderate, or severe, based on the SFMPQ scores. Moreover, the pain categories were further collapsed into a binary system consisting of none to mild, and moderate to severe, and these categories were correlated with and plotted against the preoperative and postoperative radiographic alignment categories (described above). Operative Technique and Pedal Realignment The surgical technique used in the patients described in this article entailed a standard 2-incision approach, with 1 lateral and 1 medial incision located over the target joints. Appropriate joint preparation was achieved using a combination of a resection, curettage, fish scaling, and subchondral drilling. In severely deformed or posttraumatic cases, wedged joint resection was usually unavoidable and often required tricortical corticocancellous bone grafting (allogeneic or autogenous) to achieve the desired degree of structural realignment. Subtalar fixation was performed from a plantar to dorsal approach using 1 or 2 large cannulated screws, and the talonavicular and calcaneocuboid fusions were typically fixated using either intermediate or large diameter cannulated screws. During the operations, fluoroscopy was used to assess the intraoperative alignment of the forefoot to the hindfoot in the transverse and frontal planes, and realignment of the subtalar joint was also assessed fluoroscopically before and after placement of the final fixation. The calcaneal axial fluoroscopic view was used to ensure that the long axis of the calcaneus was parallel to the mid-diaphyseal line of the distal tibia, before insertion of the guide wire for cannulated screw placement, which also served as provisional fixation for the talocalcaneal arthrodesis. Before provisional fixation of the talonavicular fusion interface, the forefoot was realigned parallel to the hindfoot in the transverse plane using the AP talo-first metatarsal angle, and perpendicular to the hindfoot in the frontal plane using the lateral talo-first metatarsal angle. Following layered wound closure and bandaging, the operated lower extremity was supported in a below-the-knee, non–weight bearing cast for approximately 8 to 10 weeks,
FIGURE 1 Preoperative and postoperative anteroposterior talofirst metatarsal angle.
FIGURE 2 Preoperative and postoperative anteroposterior talocalcaneal angle.
Statistical Plan
FIGURE 3 Preoperative and postoperative lateral talo-first metatarsal angle.
after which they were transitioned to partial and then to full weight bearing in a fracture brace over the ensuing month. Return to normal shoe gear, as a rule, was anticipated at approximately 3.5 months following surgery. Serial radiographs were obtained during the postoperative course to evaluate the progress of union at the arthrodesis sites, and the findings observed in these images were used to determine the patient’s weight-bearing status during the rehabilitation period. Edema, local skin temperature, and pain were other factors that were taken into consideration when the surgeon contemplated the patients’ return to weight-bearing activity.
The data were stored in a spreadsheet (Microsoft Excel 2004 for Mac, Version 11.3.7, Microsoft Corporation, Redmond, WA), and then imported into a statistical software program (Stata/SE 9.2 for Macintosh 2007, Stata Corporation, College Station, TX) for the analyses. We paid particular attention to the type of data (continuous or categorical) and whether or not the data were distributed in a normal or non-normal fashion; and based on these determinations, we used compatible statistical analyses when we described the data and when we compared the preoperative to postoperative radiographic and pain measurements. We also computed Spearman’s rank correlation coefficient (rho) in an effort to measure dependence between the different radiographic angles and patient outcomes in terms of postoperative pain. Values for Spearman’s rho range from 1 (100% negative association, or perfect inversion) to +1 (100% positive association, or perfect agreement), and a value of 0 indicates the absence of association. Values were also calculated for the probability that any of the radiographic measurements were independent of the likelihood that a patient would experience an outcome of mild or moderate postoperative pain. For all of the analytical calculations, statistical significance was defined at the 5% (P # .05) level. Results Inspection revealed our data to be distributed in a nonnormal fashion, hence nonparametric statistical methods were used throughout the analyses. Statistical descriptions of the data are depicted in Tables 1 and 2. The main indication for the operation was the presence of marked pain that was unresponsive to nonoperative treatment, and that inhibited daily weight-bearing activities. For the 28 fusions VOLUME 48, NUMBER 3, MAY/JUNE 2009
287
FIGURE 4 Preoperative and postoperative calcaneal inclination angle.
TABLE 1 Descriptive statistics for the continuous numeric variables comprising the dataset (N = 28 triple arthrodeses in 24 patients) Variable
Mean SD
Age at the time of surgery, y AP talocalcaneal angle ( ) Preop Postop AP talo-first metatarsal angle ( ) Preop Postop Calcaneal inclination angle ( ) Preop Postop Lateral talocalcaneal angle ( ) Preop Postop Lateral talo-first metatarsal angle ( ) Preop Postop Talar declination angle ( ) Preop Postop Duration of postop non– weight bearing, wk Time to osseous union, wk
46.75 18.28
54.5 (14, 74)
24.92 9.81 18.47 6.7
26 (10.3, 47) 18.95 (6.5, 31.6)
17.22 8.8 8.73 5.97
16.6 (2.1, 36.5) 8.6 (0.2, 25.2)
Median (Minimum, Maximum)
14.20 5.64 16.34 5.41
14.4 (1.1, 28) 15.65 (5.9, 27)
40.00 9.61 41.05 11.36
42.2 (22.3, 52.7) 38.55 (25.4, 75.9)
17.63 8.10 6.54 4.82
16.2 (3.8, 34.7) 5.9 (0.5, 17.3)
25.04 11.93 21.91 4.35 9.80 3.30
25 (0.9, 49.4) 21 (15, 30.7) 10.5 (8, 15.5)
15.60 5.42
13.4 (8.5, 23.7)
AP, anteroposterior; Preop, preoperative; Postop, postoperative.
288
THE JOURNAL OF FOOT & ANKLE SURGERY
FIGURE 5 Preoperative and postoperative lateral talocalcaneal angle.
(in 24 patients), the median patient age at the time of the operation was 54.5 (range 14 to 74) years, and the median length of follow-up was 19 (range 12 to 38) months. Moreover, the median duration of postoperative non–weight bearing was 10.5 (range 8.0 to 15.5) weeks, and the median time to osseous union was 13.4 (range 8.0 to 23.5) weeks. In regard to the sex of the patients, there were 6 (25%) males in the case series. In regard to the side on which the triple arthrodesis was performed, 9 (32.1%) of the procedures were performed on the right foot. Adjunct surgical procedures were performed in 20 (71.43%) patients, and the only other osseous procedure performed was coalition resection. Achilles tendon lengthening was performed in 21 (75%) patients. A total of 84 joints were fused to effect triple arthrodesis in 28 feet, in 24 patients. As for the method of resection for each of the specific joints involved in the triple arthrodeses, cartilage debridement was used in 73 (86.9%) of the joints involved, and wedge resection of the articular surfaces was used in 11 (13.1%) of the joints, most of which were performed at the calcaneocuboid joint (Table 2). The most common form of interfragmental compression screw fixation used for osteosynthesis in this series of triple arthrodeses was two 7.3-mm screws in 21 (75%) of the talocalcaneal fusions, whereas a single 7.3-mm lag screw was used in 6 (21.43%) of the talocalcaneal fusions. In regard to fusion of the talonavicular joint, 2 lag screws, either 4.0-mm or 4.5-mm diameter, accounted for 17 (60.71%) of the fusions. And, in regard to fusion of the calcaneocuboid joint, either a single 4.5-mm
TABLE 2 Prevalence of specific arthrodesis variables (N = 28 triple arthrodeses in 24 patients)* Variable
Count (%)
Male sex Right foot Adjunct surgery performed Adjunct coalition resection Achilles lengthening Joint resection‡ Fusion site Cartilage debridement (n = 73 [86.9%]) Wedge resection (n = 11 [13.1%]) Fixation‡ Two 4.0-mm screws (n = 10 [35.71%]) Three 4.0-mm screws (n = 1 [1.19%]) One 4.5-mm screw (n = 6 [7.14%]) Two 4.5-mm screws (n = 14 [16.67%]) One 7.3-mm screw (n = 26 [30.95%]) Two 7.3-mm screws (n = 21 [25%]) One Richards staple (n = 2 [2.38]) Two Richards staples (n = 1 [1.19]) External fixator (n = 3 [3.57]) Bone graft‡,§ Autogenous (n = 6 [7.14%]) Allogeneic (n = 1 [1.19%]) Complication Radiographic delayed union (n = 1 [1.19%]) Radiographic nonunion (n = 4 [4.76%]) Any radiographic delayed or nonunion Symptomatic nonunion Chronic edema Painful hardware Wound dehiscence Infection with dehiscence Painful ankle Postoperative pain
6 (25)† 9 (32.14) 20 (71.43) 4 (14.29) 21 (75) Talocalcaneal (n = 28) 26 (30.95%) 2 (2.38%)
Talonavicular (n = 28) 26 (30.95%) 2 (2.38%)
Calcaneocuboid (n = 28) 21 (25%) 7 (8.3%)
0 0 0 0 6 (7.14%) 21 (25%) 0 0 1 (1.19%)
10 (35.71%) 1 (1.19%) 2 (2.38%) 7 (8.3%) 7 (8.3%) 0 0 0 1 (1.19%)
0 0 4 (4.76%) 7 (8.3%) 13 (15.48%) 0 2 (2.38%) 1 (1.19%) 1 (1.19%)
2 (2.38%) 0
2 (2.38%) 0
2 (2.38%) 1 (1.19%)
0 1 (1.19%)‡ 0 0 1 (1.19%)‡ 3 (3.57%)‡ 5 (5.95% of fusion sites, 17.86% of cases, 20.83% of patients) 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) 5 (20.83%) 9 (37.5%) 6 (25%) 4 (16.67%) 1 (4.17%) 1 (4.17%)
*Value presented as a proportion of procedures unless otherwise specified (N = 28). †Value presented as a proportion of patients (N = 24). ‡Value presented as a proportion of joints fused (N = 84). §Iliac crest bone graft.
or 7.3-mm lag screw was used in 17 (60.71%) of the fusions. Large staples were used to fuse 3 (10.71%) of the calcaneocuboid joints, and an external fixator was used to stabilize all of the elements of the triple arthrodesis in 1 (3.57%) case. With the exception of the patient stabilized with the external fixator, all of the patients were immobilized in a short leg cast following the operation. Table 2 also shows that autogenous iliac crest bone graft was used in a total of 6 (7.14%) of the fusion sites, 2 (2.38%) each at the talocalcaneal, talonavicular, and calcaneocuboid joints, and an allogeneic iliac crest graft was used in 1 (1.19%) calcaneocuboid joint. Table 2 also shows that radiographic evidence of delayed union took place in 1 (1.19%) talonavicular fusion site, and radiographic evidence of nonunion occurred at 4 (4.76%) fusion sites, including 1 (1.19%) talonavicular and 3 (3.57%) calcaneocuboid sites. In other words, radiographic evidence of delayed or nonunion occurred in 5 (5.95% of fusion sites, 17.86% of cases, 20.83% of patients) fusion
sites. However, only 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) of the nonunions was clinically symptomatic. Furthermore, 1 (25%) of the patients who experienced an asymptomatic nonunion was a smoker. Still further, the 1 patient who experienced the symptomatic nonunion was an active cigarette smoker that was also lost to follow-up while using an external bone growth stimulator. None of the patients in the case series underwent revisional arthrodesis during the course of the observation period. Table 2 also shows that other observed complications included chronic edema in 5 (20.83%) patients, painful hardware requiring removal in 9 (37.50%) patients, wound dehiscence in 6 (25.00%) patients, infection with dehiscence in 4 (16.67%) patients, ankle pain in 1 (4.17%) patient, and prolonged postoperative pain in 1 (4.17%) patient. Of the 6 patients who experienced postoperative infection and/or wound dehiscence, 4 had their infection resolve with the use of systemic antibiotics and local wound care, whereas VOLUME 48, NUMBER 3, MAY/JUNE 2009
289
TABLE 3 Comparison of preoperative to postoperative radiographic angles (N = 28 triple arthrodeses in 24 patients, median followup duration = 19 [range 12 to 38] months) Variable
Preoperative ( ), median (range)
Postoperative ( ), median (range)
P value*
Change ( ), median (range)
AP talocalcaneal angle ( ) AP talar-first metatarsal angle ( ) Calcaneal inclination angle ( ) Lateral talocalcaneal angle ( ) Lateral talar-first metatarsal angle ( ) Talar declination angle ( )
26 (10.3, 47) 16.6 (2.1, 36.5) 14.4 (1.1, 28) 42.2 (22.3, 52.7) 16.2 (3.8, 34.7) 25 (0.9, 49.4)
18.95 (6.5, 31.6) 8.6 (0.2, 25.2) 15.65 (5.9, 27) 38.55 (25.4, 75.9) 5.9 (0.5, 17.3) 21 (15, 30.7)
.001 .0001 .0612 .9772 <.0001 .2355
6 ( 11.4, 27.5) 6.7 ( 8, 23) 1.55 ( 9.3, 6.7)† 2 ( 34.8, 20.4) 10.4 (1, 26.2) 5 ( 19.4, 21.5)
AP, anteroposterior. *Wilcoxon signed ranks test. †A negative change in the calcaneal inclination angle represents an increase in calcaneal inclination. TABLE 4 Comparison of preoperative to postoperative shortform McGill visual analog pain scale subjective pain assessment (N = 18 procedures in 18 patients) Description of Pain
Preoperative
None Mild Moderate Severe
0 0 9 (50.00%) 9 (50.00%)
Postoperative 5 (27.78%) 9 (50.00%) 3 (16.67%) 1 (5.56%)
P value*
.028
*c2 test for trend across ordered groups.
the other 2 required surgical debridement of their infected postoperative sites. Of the 2 patients who required surgical debridement, 1 eventually required application of a split thickness skin graft to achieve wound coverage. All 6 of these patients went on to full healing. The results of the null hypothesis (H0: there is no difference between the preoperative and postoperative measurements) tests used to compare the preoperative to postoperative radiographic alignments are depicted in Table 3. The median AP talocalcaneal angle was 26.0 (range 10.3 to 47.0 ), whereas the median postoperative AP talocalcaneal angle was 18.95 (range 6.5 to 31.6 ), and this difference was statistically significant (P = .001). The median AP talo-first metatarsal angle was16.6 (range 2.1 to 36.5 ), whereas the median postoperative AP talo-first metatarsal angle was 8.6 (range 0.2 to 25.2 ), and this difference was also statistically significant (P = .0001). The median preoperative calcaneal inclination angle was 14.4 (range 1.1 to 28.0 ), whereas this angle in the postoperative period was 15.65 (range 5.9 to 27.0 ), and this difference was not statistically significant (P = .0612). The median preoperative lateral talocalcaneal angle was 42.2 (range 22.3 to 52.7 ), whereas this angle in the postoperative period was 38.55 (range 25.4 to 75.9 ), and this difference was not statistically significant (P = .9772). The median preoperative lateral talo-first metatarsal angle was 16.2 (range 3.8 to 34.7 ), and in the postoperative period this angle measured 5.9 (range 0.5 to 17.3 ), and this difference was statistically significant (P < .0001). The median preoperative talar declination angle was 25.0 (range 290
THE JOURNAL OF FOOT & ANKLE SURGERY
0.9 to 49.4 ), whereas the median postoperative talar declination angle was 21.0 (range 15.0 to 30.7 ), and this difference was not statistically significant (P = .2355). Overall, triple arthrodesis reduced the value of all of the radiographic variables, with the exception of the calcaneal inclination angle, although only the reductions observed in the AP talocalcaneal, AP talo-first metatarsal, and lateral talo-first metatarsal angles were statistically significant (Table 3). The short-form McGill visual analog pain scale questionnaire (SF-MPQ) was used to evaluate the preoperative and postoperative pain in 18 of the patients, because not all of the 24 patients returned for the long-term postoperative assessment; these results are depicted in Table 4. Nine (50%) patients had severe pain and 9 (50%) had moderate pain preoperatively; whereas, in the postoperative period, 1 (5.56%) patient had severe pain, 3 (16.67%) had moderate pain, 9 (50%) had mild pain, and 5 (27.78%) had no pain; these differences were statistically significant (P = .028). For all of the radiographic variables and pain outcomes, we performed correlation analyses using Spearman’s rank correlation coefficient in an effort to determine whether or not any of the outcomes were dependent on the radiographic measurements. Of these computations, only the correlation of none to mild postoperative pain with a preoperative lateral talocalcaneal angle equal to 35 to 50 , and the correlation of none to mild postoperative pain with a preoperative talar declination angle of 21 or less, were statistically significant (P = .0223 and P = .0095, respectively). Figure 6 shows that triple arthrodesis realigned the foot such that the likelihood of experiencing none to mild postoperative pain decreased (inverse correlation) if the preoperative lateral talocalcaneal angle was normal. Figure 7 shows that triple arthrodesis realigned the foot such that the likelihood of experiencing none to mild postoperative pain increased if the preoperative lateral talar declination angle was normal. Discussion A successful long-term outcome following triple arthrodesis is characterized by a return to normal shoe gear and
1.5
1.5
1
1
.5
.5
0
0
-.5
-.5 0
1
2
3
pre_tc_lat_norm Fitted values
0
1
2
3
postpain
postpain 95% CI
pre_talar_dec_norm Fitted values
95% CI
FIGURE 6 Scatter plot of preoperative lateral talocalcaneal angle = 35 to 50 versus postoperative pain = mild or none (Spearman’s rho = 0.5497, probability of independence = 0.0223). This plot shows that a normal (35 to 50 ) preoperative lateral talocalcaneal angle was moderately inversely correlated (not likely to be independent) with none to mild postoperative pain.
FIGURE 7 Scatter plot of preoperative talar declination angle # 21 versus postoperative pain = mild or none (Spearman’s rho = 0.6261, probability of independence = 0.0095). This plot shows that a normal (# 21 ) preoperative lateral talar declination angle was strongly positively correlated (not likely to be independent) with none to mild postoperative pain.
a functional ambulatory status with no significant gait disturbances, and minimal to no pain. Triple arthrodesis has been shown to yield successful results in greater than 90% of cases (1–2). Historically, nonunion has been implicated as a common cause of failure, and rates of nonunion following triple arthrodesis have been reported to range from about 5% to 23% (2, 4, 7, 13). Lateral column overload and instability, stiffness, and progression of degenerative changes in adjacent joints are other potential complications related to the triple arthrodesis, and all of these complications can be attributed to misalignment of the hindfoot fusions. Triple arthrodesis is used for end-stage hindfoot arthrosis caused by a wide range of etiologies, and severe varus and valgus deformation are common deformities treated with this form of reconstruction. One report, in fact, described a series of 132 feet in which 102 displayed a valgus deformity preoperatively, whereas 13 displayed varus deformity that required surgical repair; and these investigators found that there was a significant association between patient satisfaction and postoperative alignment (2). And, although nonunion is an often cited complication of triple arthrodesis, malposition is another potential factor that can compromise a successful long-term outcome. In an analysis of triple arthrodesis, misalignment of the fusion was noted to account for 66% of the failures (8). In another report of a series of 17 adults who underwent triple arthrodesis, 6 of the feet were noted to display moderate to severe valgus position of the heel following the surgery, and 4 of these developed ankle degenerative joint disease and 3 experienced persistent difficulty wearing shoes because of persistent valgus deformation of the hindfoot (6). The authors of the current report have, for several years, used a protocol for evaluating the alignment achieved by
the triple arthrodesis (14), and the purpose of the current investigation was to review subjective (pain) and objective (radiographic) outcomes in 24 patients (28 feet) who underwent a triple arthrodesis when these specific realignment considerations were used. It was interesting to note that triple arthrodesis led to what would be considered improvement in all of the radiographic variables, although only the changes in the AP talocalcaneal, AP talo-first metatarsal, and lateral talofirst metatarsal angles were statistically significant (Table 2). These observed changes attest to the power that the surgeon has in regard to altering the structural alignment of the foot when undertaking triple arthrodesis. Specifically, the data in the current study demonstrated a median change of 10.4 (range 1.0 to 26.2 ) for the lateral talo-first metatarsal angle, whereas a median change of 6.7 (range 8.0 to 23.0 ) was observed for the AP talo-first metatarsal angle (Table 3). It appears from these results that the average realignment of the forefoot to the hindfoot in the sagittal plane was greater than realignment in the transverse plane. These results are comparable to other reports of angular correction in the frontal and sagittal planes (2, 6). Moreover, the AP talocalcaneal angle displayed a median change of 6.86 (range 11.40 to 27.50 ), compared with a median 2.0 (range 34.8 to 20.4 ) change measured for the lateral talocalcaneal angle (Table 3); and these changes suggested that realignment within the hindfoot (between the talus and calcaneus) involved more transverse plane movement of the talus back over the calcaneus, rather than markedly changing the pitch of either the talus or the calcaneus in the sagittal plane. In fact, Table 3 also shows that the median changes in the calcaneal inclination and talar declination angels were 1.55 (range 9.30 to 6.70 ) and 5.00 (range 19.40 VOLUME 48, NUMBER 3, MAY/JUNE 2009
291
to 21.50 ), respectively, suggesting that what change did occur in the sagittal plane alignment of the hindfoot occurred primarily by means of reduction of the downward pitch of the talus rather than elevation of the pitch of the calcaneus. This observation is probably consistent with the general goal of establishing a stable, plantigrade foot by means of triple arthrodesis, and is probably more in keeping with reconstruction of the pes valgus deformity rather than repair of the cavus foot. No calcaneal osteotomies were required to achieve the desired amount of tarsal realignment, although we used autogenous bone graft at 6 (7.14%) arthrodesis sites and allograft at 1 (1.19%) calcaneocuboid site, and posterior soft tissue release in the form of Achilles tendon lengthening, or tenotomy, was performed in 21 (75.00%) of the procedures (Table 2). Again, we feel that these changes, which can be used to markedly alter the relationship of the forefoot to the hindfoot, and that of the hindfoot to the leg, are related primarily to the freedom that the surgeon has in regard to selecting the optimal orientation of the tarsals following resection of the midtarsal and subtalar joints. Nonunion rates have decreased to about 5% in the recent literature (2, 4, 7, 13), and we observed approximately the same prevalence of radiographic evidence of nonunion, namely 4 (4.76%) in 28 triple arthrodeses. Interestingly, 3 (60%) of the nonunions occurred at the calcaneocuboid joint, and we feel that this is probably the result of the obliquity of interfragmental compression screw fixation across the fusion interface, and the vulnerability of this fusion site to weightbearing forces sustained during the postoperative phase. In the future, it would be interesting to see if the addition of buttressing staples, or even a neutralization plate, might substantially decrease the prevalence of nonunion at this site (or the talonavicular fusion site, for that matter). The short-version McGill visual analog pain scale questionnaire was used to evaluate the preoperative and postoperative pain for 18 of the triple arthrodeses, because the patients failed to return the questionnaire following 10 of the procedures. Overall, we observed a statistically significant decrease in pain between the preoperative and postoperative periods, and 77.78% of the patients experienced none to mild pain long-term following the operation (Table 4). We also measured dependence between the radiographic variables and pain following triple arthrodesis using Spearman’s rank correlation coefficient, which is a distribution-free test of independence between 2 variables that makes no assumptions about continuity of the data, and assumes only that the data are ordered by rank. A look at Figures 6 and 7 shows that triple arthrodesis realigned the foot such that the likelihood of experiencing none to mild postoperative pain decreased if the preoperative lateral talocalcaneal angle was normal to begin with, and it increased if the preoperative lateral talar declination angle was normal. This first observation is somewhat intuitive, in that it makes sense to think that a more normal lateral talar declination angle would require less realignment in the sagittal plane, hence 292
THE JOURNAL OF FOOT & ANKLE SURGERY
the triple arthrodesis (or any reconstructive osteotomy or arthrodesis) would be more likely to satisfactorily establish a plantigrade foot, which would likely be less painful than one that remained in a pes valgus (or varus) malposition. Although not intuitively obvious, it was interesting to note that if the lateral talocalcaneal angle was normal to begin with, it is likely that it would be more difficult (inversely correlated) to alleviate pain following realignment via triple arthrodesis. The precise meaning of this finding remains obscure, however we suspect that it has to do with the interaction of corrective motion of the calcaneus relative to the talus in both the transverse and sagittal planes. Like all observational investigations, a number of methodological limitations probably threaten the validity of the conclusions that we have drawn from our results. The fact that not all of the patients returned the pain questionnaire following the operation meant that we observed a 35.71% loss to follow-up that, in all likelihood influenced our interpretation of the alteration of pain following with triple arthrodesis. Although we are not able to clearly state why the patients did not return the pain questionnaire, hence, lost to follow-up, we did not anecdotally identify any specific reason for the loss. Furthermore, in this investigation, the operating surgeons also measured and recorded the preoperative and postoperative radiographic values, and this could have imparted some degree of bias toward an effect of the surgery. In conclusion, triple arthrodesis appears to be a powerful method of realigning the forefoot to the hindfoot. It can be used to clinically and statistically significantly improve the structural alignment of a number of radiographic measurements that describe the orientation of the tarsals, namely the AP talocalcaneal angle, the AP talo-first metatarsal angle, and the lateral talo-first metatarsal angle. Our results also suggest that triple arthrodesis realigns the relationship of the talus and calcaneus primarily by means of transverse plane movement of the talus back over the calcaneus, rather than by altering the sagittal plane pitch of the calcaneus; and that change in the sagittal plane alignment of the hindfoot takes place primarily by means of reduction of the downward pitch of the talus rather than by means of elevation of the pitch of the calcaneus. Although we observed only 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) symptomatic nonunion in this case series, this complication remains a risk whenever arthrodesis is undertaken. In the future, the results of this case series may be useful in the development of randomized controlled trials or prospective cohort studies that further explore the risk factors and outcomes associated with triple arthrodesis.
Acknowledgment We appreciate the work of D. Scot Malay, DPM, MSCE, FACFAS, who performed the statistical analyses described in this investigation.
References 1. Beischer AD, Brodsky JW, Pollo FE, Peereboom J. Functional outcome and gait analysis after triple or double arthrodesis. Foot Ankle Int 20(9): 545–553, 1999. 2. Pell RF, Myerson MS, Schon LC. Clinical outcome after primary triple arthrodesis. J Bone Joint Surg Am 82-A:47–57, 2000. 3. Angus PD, Cowell HR. Triple arthrodesis: a critical long-term review. J Bone Joint Surg Br 68-B(2):260–265, 1986. 4. Saltzman CL, Fehrle MJ, Cooper RR, Spencer EC, Ponseti IV. Triple arthrodesis: twenty-five and forty-four-year average follow-up of the same patients. J Bone Joint Surg 81-A(10):1391–1402, 1999. 5. Sangeorzan BJ, Smith D, Veith R, Hansen ST. Triple arthrodesis using internal fixation in treatment of adult foot disorders. Clin Orthop 294: 299–307, 1993. 6. Graves SC, Mann RA, Graves KO. Triple arthrodesis in older adults: results after long-term follow-up. J Bone Joint Surg Am 75-A(3):355–362, 1993. 7. Fortin PT, Walling AK. Triple arthrodesis. Clin Orthop 365:91–99, 1999.
8. Maenpaa H, Lehto MU, Belt EA. What went wrong with in triple arthrodesis. Clin Orthop 391:218–223, 2001. 9. Bednarz PA, Monroe MT, Manoli A 2nd. Triple arthrodesis in adults using rigid internal fixation: an assessment of outcome. Foot Ankle Int 20(6):356–363, 1999. 10. Wenz W, Bruckner T, Akbar M. Complete tendon transfer and inverse Lambrinudi arthrodesis: preliminary results of a new technique for the treatment of paralytic pes calcaneus. Foot Ankle Int 29(7):683–689, 2008. 11. Malay DS. The PI Manual: A Handbook of Podiatric Medicine and Surgery, ed 2, pp 214 and 237, The Podiatry Institute, Decatur, GA, 2008. 12. Melzack R. The short-form McGill Pain Questionnaire. Pain 30: 191–197, 1987. 13. Wukich DK, Bowen JR. A long-term study of triple arthrodesis for correction of pes cavovarus in Charcot-Marie-Tooth disease. J Pediatr Orthop 9(4):433–437, 1989. 14. Catanzariti AR, Mendicino RW, Whitaker JM, Reeves CL. Realignment considerations in the triple arthrodesis. J Am Podiatr Med Assoc 95(1): 13–17, 2005.
VOLUME 48, NUMBER 3, MAY/JUNE 2009
293
The Effect of Hindfoot Realignment in Triple Arthrodesis Brandon J. Child, DPM,1 Joel Hix, DPM,1 Alan R. Catanzariti, DPM, FACFAS,2 Robert W. Mendicino, DPM, FACFAS,3 and Karl Saltrick, DPM, FACFAS4 Triple arthrodesis is a useful form of surgical intervention for repair of advanced hindfoot deformity. Crucial to the success of triple arthrodesis is achievement of a suitable alignment of the hindfoot in relation to the leg and forefoot. A number of radiographic measurements can be used to describe the resultant alignment of the foot and, in this article, we present the results of a review of a series of 28 triple arthrodeses, in 24 patients, followed for a median duration of 19 (range 12 to 38) months. The results of this investigation showed clinically and statistically significant improvements in the anteroposterior talocalcaneal angle and talo-first metatarsal angles, and the lateral talo-first metatarsal angle; and triple arthrodesis realigned the hindfoot primarily by means of transverse plane movement of the talus over the calcaneus, and sagittal plane realignment occurred primarily by means of reduction of the downward pitch of the talus rather than by means of elevation of the pitch of the calcaneus. Only 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) symptomatic nonunion was observed. Correlation analyses showed that a normal (35 to 50 ) preoperative lateral talocalcaneal angle was moderately inversely correlated, and a normal (# 21 ) preoperative lateral talar declination angle was strongly positively correlated, with none to mild postoperative pain. Based on these findings, we concluded that triple arthrodesis satisfactorily realigned the hindfoot and reduced pain in patients with advanced arthrosis. Level of Clinical Evidence: 4 (The Journal of Foot & Ankle Surgery 48(3):285–293, 2009) Key Words: arthrosis, cavus, deformity, fixation, pes valgus, varus
The primary goals of triple arthrodesis are to provide stability, decrease pain, and reduce deformity; and appropriate realignment of the hindfoot is a key factor in achieving an acceptable surgical outcome. A well-aligned triple arthrodesis should ultimately result in a plantigrade foot that will support the ankle and reduce the risk of complications such as weight-bearing instability, the progression of Address correspondence to: Robert W. Mendicino, DPM, FACFAS, The Foot and Ankle Institute of Western Pennsylvania, 4800 Friendship Avenue, North Tower, First Floor, Pittsburgh, PA 15224. E-mail: rwmendicino@ faiwp.com 1 Third-Year Resident, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. 2 Director of Residency Training, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. 3 Chief, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. 4 Assistant Director of Residency Training, Department of Foot and Ankle Surgery, The Western Pennsylvania Hospital, Pittsburgh, PA. Financial Disclosure: None reported. Conflict of Interest: None reported. Copyright Ó 2009 by the American College of Foot and Ankle Surgeons 1067-2516/09/4803-0002$36.00/0 doi:10.1053/j.jfas.2009.02.006
foot and/or ankle deformity, and the development of secondary degeneration of adjacent joints. Although triple arthrodesis is generally known to result in favorable longterm results relative to pain relief and patient satisfaction (1–5), we feel that there has been limited attention paid to the importance of precise hindfoot realignment as it relates to the outcome of the procedure. Furthermore, poor outcomes have been reported in association with tarsal misalignment, such as fixed hindfoot varus (6–8). Although it is widely accepted that surgeons should avoid fusion of the hindfoot in varus, a thorough review of the biomedical databases revealed just a few reports (5, 9, 10) that quantitatively described the association of tarsal realignment with the outcome of a triple arthrodesis. Sangeorzan and colleagues (5) observed 44 triple arthrodeses in 40 patients, with an average of 4.9 years of follow-up, and noted an average improvement of 17 in the lateral talometatarsal angle, 11 in the lateral talocalcaneal angle, and 18 in the anteroposterior (AP) talometatarsal angle. Bednarz et al (9) observed 63 triple arthrodeses in 57 patients, at an average of 30 months follow-up, and noted an average improvement of 12 in the VOLUME 48, NUMBER 3, MAY/JUNE 2009
285
lateral talometatarsal angle, 7 in the lateral talocalcaneal angle, and 10 in the AP talometatarsal angle (P < .0001). Finally, Wenz et al (10) observed 6 patients with paralytic pes calcaneus at an average follow-up of 32 months following triple arthrodesis, and noted statistically significant improvements in the talocalcaneal angle, the calcaneal pitch, and the talo-first metatarsal angle. The purpose of the current investigation was to review subjective and objective outcomes in 24 patients who underwent 28 triple arthrodeses, with particular emphasis paid to the influence that hindfoot alignment had on the long-term measurement of radiographic angles and reduction of pain following the operation. Patients and Methods Chart Review The authors reviewed the records of 24 patients who had undergone triple arthrodesis (28 procedures) at the Foot and Ankle Institute of Western Pennsylvania, at The Western Pennsylvania Hospital, in Pittsburgh, Pennsylvania, during the period extending from September 2000 to May 2003. After obtaining approval from the Institutional Review Board, patient information was abstracted from the patient records and radiographic measurements were made using the preoperative and postoperative radiographic images, and a long-term follow-up pain questionnaire was dispensed and returned via the postal service. The questionnaire was sent to all patients who underwent triple arthrodesis during the above captioned time period. Radiographs were measured at a median duration of 19 months. Radiographic Assessment Preoperative and postoperative radiographs were compared to determine the degree of structural realignment of the foot. To be included in the case series, patients had to have at least 1-year follow-up notes and radiographs in the medical record. To assess forefoot-to-hindfoot relationships, AP talo-first metatarsal and talocalcaneal angles were used to determine the transverse plane alignment (Figures 1 and 2). On the lateral foot radiograph, the talo-first metatarsal, talocalcaneal, calcaneal inclination, and talar declination angles were measured (Figures 3–5). Furthermore, to correlate radiographic alignment with pain outcomes, the radiographic measurements were categorized relative to normal values (11) as follows: AP talocalcaneal angle 25 or less, AP talofirst metatarsal angle less than 15 , calcaneal inclination angle equals 15 to 21 , lateral talocalcaneal angle equals 35 to 50 , lateral talo-first metatarsal angle equals 5 to 5 , and talar declination angle 21 or less. Still further, the time to successful osseous union was also determined radiographically by assessing serial postoperative radiographs, and successful union 286
THE JOURNAL OF FOOT & ANKLE SURGERY
was considered present when there was evidence of trabecular bone crossing the arthrodesis sites. Pain Assessment The short-form McGill visual analog pain scale questionnaire (SF-MPQ) (12) was used to evaluate preoperative and postoperative pain. The SF-MPQ has 15 pain descriptors that are subjectively rated by the patient, and uses an intensity scale of 0 (none) to 3 (severe), with the minimum score being 15 (no pain) and the maximum score being 45 (severe pain). The SF-MPQ has been found to be reliable and to provide valid information related to pain in acute and chronic situations (12). Preoperative and postoperative pain was also categorized as none, mild, moderate, or severe, based on the SFMPQ scores. Moreover, the pain categories were further collapsed into a binary system consisting of none to mild, and moderate to severe, and these categories were correlated with and plotted against the preoperative and postoperative radiographic alignment categories (described above). Operative Technique and Pedal Realignment The surgical technique used in the patients described in this article entailed a standard 2-incision approach, with 1 lateral and 1 medial incision located over the target joints. Appropriate joint preparation was achieved using a combination of a resection, curettage, fish scaling, and subchondral drilling. In severely deformed or posttraumatic cases, wedged joint resection was usually unavoidable and often required tricortical corticocancellous bone grafting (allogeneic or autogenous) to achieve the desired degree of structural realignment. Subtalar fixation was performed from a plantar to dorsal approach using 1 or 2 large cannulated screws, and the talonavicular and calcaneocuboid fusions were typically fixated using either intermediate or large diameter cannulated screws. During the operations, fluoroscopy was used to assess the intraoperative alignment of the forefoot to the hindfoot in the transverse and frontal planes, and realignment of the subtalar joint was also assessed fluoroscopically before and after placement of the final fixation. The calcaneal axial fluoroscopic view was used to ensure that the long axis of the calcaneus was parallel to the mid-diaphyseal line of the distal tibia, before insertion of the guide wire for cannulated screw placement, which also served as provisional fixation for the talocalcaneal arthrodesis. Before provisional fixation of the talonavicular fusion interface, the forefoot was realigned parallel to the hindfoot in the transverse plane using the AP talo-first metatarsal angle, and perpendicular to the hindfoot in the frontal plane using the lateral talo-first metatarsal angle. Following layered wound closure and bandaging, the operated lower extremity was supported in a below-the-knee, non–weight bearing cast for approximately 8 to 10 weeks,
FIGURE 1 Preoperative and postoperative anteroposterior talofirst metatarsal angle.
FIGURE 2 Preoperative and postoperative anteroposterior talocalcaneal angle.
Statistical Plan
FIGURE 3 Preoperative and postoperative lateral talo-first metatarsal angle.
after which they were transitioned to partial and then to full weight bearing in a fracture brace over the ensuing month. Return to normal shoe gear, as a rule, was anticipated at approximately 3.5 months following surgery. Serial radiographs were obtained during the postoperative course to evaluate the progress of union at the arthrodesis sites, and the findings observed in these images were used to determine the patient’s weight-bearing status during the rehabilitation period. Edema, local skin temperature, and pain were other factors that were taken into consideration when the surgeon contemplated the patients’ return to weight-bearing activity.
The data were stored in a spreadsheet (Microsoft Excel 2004 for Mac, Version 11.3.7, Microsoft Corporation, Redmond, WA), and then imported into a statistical software program (Stata/SE 9.2 for Macintosh 2007, Stata Corporation, College Station, TX) for the analyses. We paid particular attention to the type of data (continuous or categorical) and whether or not the data were distributed in a normal or non-normal fashion; and based on these determinations, we used compatible statistical analyses when we described the data and when we compared the preoperative to postoperative radiographic and pain measurements. We also computed Spearman’s rank correlation coefficient (rho) in an effort to measure dependence between the different radiographic angles and patient outcomes in terms of postoperative pain. Values for Spearman’s rho range from 1 (100% negative association, or perfect inversion) to +1 (100% positive association, or perfect agreement), and a value of 0 indicates the absence of association. Values were also calculated for the probability that any of the radiographic measurements were independent of the likelihood that a patient would experience an outcome of mild or moderate postoperative pain. For all of the analytical calculations, statistical significance was defined at the 5% (P # .05) level. Results Inspection revealed our data to be distributed in a nonnormal fashion, hence nonparametric statistical methods were used throughout the analyses. Statistical descriptions of the data are depicted in Tables 1 and 2. The main indication for the operation was the presence of marked pain that was unresponsive to nonoperative treatment, and that inhibited daily weight-bearing activities. For the 28 fusions VOLUME 48, NUMBER 3, MAY/JUNE 2009
287
FIGURE 4 Preoperative and postoperative calcaneal inclination angle.
TABLE 1 Descriptive statistics for the continuous numeric variables comprising the dataset (N = 28 triple arthrodeses in 24 patients) Variable
Mean SD
Age at the time of surgery, y AP talocalcaneal angle ( ) Preop Postop AP talo-first metatarsal angle ( ) Preop Postop Calcaneal inclination angle ( ) Preop Postop Lateral talocalcaneal angle ( ) Preop Postop Lateral talo-first metatarsal angle ( ) Preop Postop Talar declination angle ( ) Preop Postop Duration of postop non– weight bearing, wk Time to osseous union, wk
46.75 18.28
54.5 (14, 74)
24.92 9.81 18.47 6.7
26 (10.3, 47) 18.95 (6.5, 31.6)
17.22 8.8 8.73 5.97
16.6 (2.1, 36.5) 8.6 (0.2, 25.2)
Median (Minimum, Maximum)
14.20 5.64 16.34 5.41
14.4 (1.1, 28) 15.65 (5.9, 27)
40.00 9.61 41.05 11.36
42.2 (22.3, 52.7) 38.55 (25.4, 75.9)
17.63 8.10 6.54 4.82
16.2 (3.8, 34.7) 5.9 (0.5, 17.3)
25.04 11.93 21.91 4.35 9.80 3.30
25 (0.9, 49.4) 21 (15, 30.7) 10.5 (8, 15.5)
15.60 5.42
13.4 (8.5, 23.7)
AP, anteroposterior; Preop, preoperative; Postop, postoperative.
288
THE JOURNAL OF FOOT & ANKLE SURGERY
FIGURE 5 Preoperative and postoperative lateral talocalcaneal angle.
(in 24 patients), the median patient age at the time of the operation was 54.5 (range 14 to 74) years, and the median length of follow-up was 19 (range 12 to 38) months. Moreover, the median duration of postoperative non–weight bearing was 10.5 (range 8.0 to 15.5) weeks, and the median time to osseous union was 13.4 (range 8.0 to 23.5) weeks. In regard to the sex of the patients, there were 6 (25%) males in the case series. In regard to the side on which the triple arthrodesis was performed, 9 (32.1%) of the procedures were performed on the right foot. Adjunct surgical procedures were performed in 20 (71.43%) patients, and the only other osseous procedure performed was coalition resection. Achilles tendon lengthening was performed in 21 (75%) patients. A total of 84 joints were fused to effect triple arthrodesis in 28 feet, in 24 patients. As for the method of resection for each of the specific joints involved in the triple arthrodeses, cartilage debridement was used in 73 (86.9%) of the joints involved, and wedge resection of the articular surfaces was used in 11 (13.1%) of the joints, most of which were performed at the calcaneocuboid joint (Table 2). The most common form of interfragmental compression screw fixation used for osteosynthesis in this series of triple arthrodeses was two 7.3-mm screws in 21 (75%) of the talocalcaneal fusions, whereas a single 7.3-mm lag screw was used in 6 (21.43%) of the talocalcaneal fusions. In regard to fusion of the talonavicular joint, 2 lag screws, either 4.0-mm or 4.5-mm diameter, accounted for 17 (60.71%) of the fusions. And, in regard to fusion of the calcaneocuboid joint, either a single 4.5-mm
TABLE 2 Prevalence of specific arthrodesis variables (N = 28 triple arthrodeses in 24 patients)* Variable
Count (%)
Male sex Right foot Adjunct surgery performed Adjunct coalition resection Achilles lengthening Joint resection‡ Fusion site Cartilage debridement (n = 73 [86.9%]) Wedge resection (n = 11 [13.1%]) Fixation‡ Two 4.0-mm screws (n = 10 [35.71%]) Three 4.0-mm screws (n = 1 [1.19%]) One 4.5-mm screw (n = 6 [7.14%]) Two 4.5-mm screws (n = 14 [16.67%]) One 7.3-mm screw (n = 26 [30.95%]) Two 7.3-mm screws (n = 21 [25%]) One Richards staple (n = 2 [2.38]) Two Richards staples (n = 1 [1.19]) External fixator (n = 3 [3.57]) Bone graft‡,§ Autogenous (n = 6 [7.14%]) Allogeneic (n = 1 [1.19%]) Complication Radiographic delayed union (n = 1 [1.19%]) Radiographic nonunion (n = 4 [4.76%]) Any radiographic delayed or nonunion Symptomatic nonunion Chronic edema Painful hardware Wound dehiscence Infection with dehiscence Painful ankle Postoperative pain
6 (25)† 9 (32.14) 20 (71.43) 4 (14.29) 21 (75) Talocalcaneal (n = 28) 26 (30.95%) 2 (2.38%)
Talonavicular (n = 28) 26 (30.95%) 2 (2.38%)
Calcaneocuboid (n = 28) 21 (25%) 7 (8.3%)
0 0 0 0 6 (7.14%) 21 (25%) 0 0 1 (1.19%)
10 (35.71%) 1 (1.19%) 2 (2.38%) 7 (8.3%) 7 (8.3%) 0 0 0 1 (1.19%)
0 0 4 (4.76%) 7 (8.3%) 13 (15.48%) 0 2 (2.38%) 1 (1.19%) 1 (1.19%)
2 (2.38%) 0
2 (2.38%) 0
2 (2.38%) 1 (1.19%)
0 1 (1.19%)‡ 0 0 1 (1.19%)‡ 3 (3.57%)‡ 5 (5.95% of fusion sites, 17.86% of cases, 20.83% of patients) 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) 5 (20.83%) 9 (37.5%) 6 (25%) 4 (16.67%) 1 (4.17%) 1 (4.17%)
*Value presented as a proportion of procedures unless otherwise specified (N = 28). †Value presented as a proportion of patients (N = 24). ‡Value presented as a proportion of joints fused (N = 84). §Iliac crest bone graft.
or 7.3-mm lag screw was used in 17 (60.71%) of the fusions. Large staples were used to fuse 3 (10.71%) of the calcaneocuboid joints, and an external fixator was used to stabilize all of the elements of the triple arthrodesis in 1 (3.57%) case. With the exception of the patient stabilized with the external fixator, all of the patients were immobilized in a short leg cast following the operation. Table 2 also shows that autogenous iliac crest bone graft was used in a total of 6 (7.14%) of the fusion sites, 2 (2.38%) each at the talocalcaneal, talonavicular, and calcaneocuboid joints, and an allogeneic iliac crest graft was used in 1 (1.19%) calcaneocuboid joint. Table 2 also shows that radiographic evidence of delayed union took place in 1 (1.19%) talonavicular fusion site, and radiographic evidence of nonunion occurred at 4 (4.76%) fusion sites, including 1 (1.19%) talonavicular and 3 (3.57%) calcaneocuboid sites. In other words, radiographic evidence of delayed or nonunion occurred in 5 (5.95% of fusion sites, 17.86% of cases, 20.83% of patients) fusion
sites. However, only 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) of the nonunions was clinically symptomatic. Furthermore, 1 (25%) of the patients who experienced an asymptomatic nonunion was a smoker. Still further, the 1 patient who experienced the symptomatic nonunion was an active cigarette smoker that was also lost to follow-up while using an external bone growth stimulator. None of the patients in the case series underwent revisional arthrodesis during the course of the observation period. Table 2 also shows that other observed complications included chronic edema in 5 (20.83%) patients, painful hardware requiring removal in 9 (37.50%) patients, wound dehiscence in 6 (25.00%) patients, infection with dehiscence in 4 (16.67%) patients, ankle pain in 1 (4.17%) patient, and prolonged postoperative pain in 1 (4.17%) patient. Of the 6 patients who experienced postoperative infection and/or wound dehiscence, 4 had their infection resolve with the use of systemic antibiotics and local wound care, whereas VOLUME 48, NUMBER 3, MAY/JUNE 2009
289
TABLE 3 Comparison of preoperative to postoperative radiographic angles (N = 28 triple arthrodeses in 24 patients, median followup duration = 19 [range 12 to 38] months) Variable
Preoperative ( ), median (range)
Postoperative ( ), median (range)
P value*
Change ( ), median (range)
AP talocalcaneal angle ( ) AP talar-first metatarsal angle ( ) Calcaneal inclination angle ( ) Lateral talocalcaneal angle ( ) Lateral talar-first metatarsal angle ( ) Talar declination angle ( )
26 (10.3, 47) 16.6 (2.1, 36.5) 14.4 (1.1, 28) 42.2 (22.3, 52.7) 16.2 (3.8, 34.7) 25 (0.9, 49.4)
18.95 (6.5, 31.6) 8.6 (0.2, 25.2) 15.65 (5.9, 27) 38.55 (25.4, 75.9) 5.9 (0.5, 17.3) 21 (15, 30.7)
.001 .0001 .0612 .9772 <.0001 .2355
6 ( 11.4, 27.5) 6.7 ( 8, 23) 1.55 ( 9.3, 6.7)† 2 ( 34.8, 20.4) 10.4 (1, 26.2) 5 ( 19.4, 21.5)
AP, anteroposterior. *Wilcoxon signed ranks test. †A negative change in the calcaneal inclination angle represents an increase in calcaneal inclination. TABLE 4 Comparison of preoperative to postoperative shortform McGill visual analog pain scale subjective pain assessment (N = 18 procedures in 18 patients) Description of Pain
Preoperative
None Mild Moderate Severe
0 0 9 (50.00%) 9 (50.00%)
Postoperative 5 (27.78%) 9 (50.00%) 3 (16.67%) 1 (5.56%)
P value*
.028
*c2 test for trend across ordered groups.
the other 2 required surgical debridement of their infected postoperative sites. Of the 2 patients who required surgical debridement, 1 eventually required application of a split thickness skin graft to achieve wound coverage. All 6 of these patients went on to full healing. The results of the null hypothesis (H0: there is no difference between the preoperative and postoperative measurements) tests used to compare the preoperative to postoperative radiographic alignments are depicted in Table 3. The median AP talocalcaneal angle was 26.0 (range 10.3 to 47.0 ), whereas the median postoperative AP talocalcaneal angle was 18.95 (range 6.5 to 31.6 ), and this difference was statistically significant (P = .001). The median AP talo-first metatarsal angle was16.6 (range 2.1 to 36.5 ), whereas the median postoperative AP talo-first metatarsal angle was 8.6 (range 0.2 to 25.2 ), and this difference was also statistically significant (P = .0001). The median preoperative calcaneal inclination angle was 14.4 (range 1.1 to 28.0 ), whereas this angle in the postoperative period was 15.65 (range 5.9 to 27.0 ), and this difference was not statistically significant (P = .0612). The median preoperative lateral talocalcaneal angle was 42.2 (range 22.3 to 52.7 ), whereas this angle in the postoperative period was 38.55 (range 25.4 to 75.9 ), and this difference was not statistically significant (P = .9772). The median preoperative lateral talo-first metatarsal angle was 16.2 (range 3.8 to 34.7 ), and in the postoperative period this angle measured 5.9 (range 0.5 to 17.3 ), and this difference was statistically significant (P < .0001). The median preoperative talar declination angle was 25.0 (range 290
THE JOURNAL OF FOOT & ANKLE SURGERY
0.9 to 49.4 ), whereas the median postoperative talar declination angle was 21.0 (range 15.0 to 30.7 ), and this difference was not statistically significant (P = .2355). Overall, triple arthrodesis reduced the value of all of the radiographic variables, with the exception of the calcaneal inclination angle, although only the reductions observed in the AP talocalcaneal, AP talo-first metatarsal, and lateral talo-first metatarsal angles were statistically significant (Table 3). The short-form McGill visual analog pain scale questionnaire (SF-MPQ) was used to evaluate the preoperative and postoperative pain in 18 of the patients, because not all of the 24 patients returned for the long-term postoperative assessment; these results are depicted in Table 4. Nine (50%) patients had severe pain and 9 (50%) had moderate pain preoperatively; whereas, in the postoperative period, 1 (5.56%) patient had severe pain, 3 (16.67%) had moderate pain, 9 (50%) had mild pain, and 5 (27.78%) had no pain; these differences were statistically significant (P = .028). For all of the radiographic variables and pain outcomes, we performed correlation analyses using Spearman’s rank correlation coefficient in an effort to determine whether or not any of the outcomes were dependent on the radiographic measurements. Of these computations, only the correlation of none to mild postoperative pain with a preoperative lateral talocalcaneal angle equal to 35 to 50 , and the correlation of none to mild postoperative pain with a preoperative talar declination angle of 21 or less, were statistically significant (P = .0223 and P = .0095, respectively). Figure 6 shows that triple arthrodesis realigned the foot such that the likelihood of experiencing none to mild postoperative pain decreased (inverse correlation) if the preoperative lateral talocalcaneal angle was normal. Figure 7 shows that triple arthrodesis realigned the foot such that the likelihood of experiencing none to mild postoperative pain increased if the preoperative lateral talar declination angle was normal. Discussion A successful long-term outcome following triple arthrodesis is characterized by a return to normal shoe gear and
1.5
1.5
1
1
.5
.5
0
0
-.5
-.5 0
1
2
3
pre_tc_lat_norm Fitted values
0
1
2
3
postpain
postpain 95% CI
pre_talar_dec_norm Fitted values
95% CI
FIGURE 6 Scatter plot of preoperative lateral talocalcaneal angle = 35 to 50 versus postoperative pain = mild or none (Spearman’s rho = 0.5497, probability of independence = 0.0223). This plot shows that a normal (35 to 50 ) preoperative lateral talocalcaneal angle was moderately inversely correlated (not likely to be independent) with none to mild postoperative pain.
FIGURE 7 Scatter plot of preoperative talar declination angle # 21 versus postoperative pain = mild or none (Spearman’s rho = 0.6261, probability of independence = 0.0095). This plot shows that a normal (# 21 ) preoperative lateral talar declination angle was strongly positively correlated (not likely to be independent) with none to mild postoperative pain.
a functional ambulatory status with no significant gait disturbances, and minimal to no pain. Triple arthrodesis has been shown to yield successful results in greater than 90% of cases (1–2). Historically, nonunion has been implicated as a common cause of failure, and rates of nonunion following triple arthrodesis have been reported to range from about 5% to 23% (2, 4, 7, 13). Lateral column overload and instability, stiffness, and progression of degenerative changes in adjacent joints are other potential complications related to the triple arthrodesis, and all of these complications can be attributed to misalignment of the hindfoot fusions. Triple arthrodesis is used for end-stage hindfoot arthrosis caused by a wide range of etiologies, and severe varus and valgus deformation are common deformities treated with this form of reconstruction. One report, in fact, described a series of 132 feet in which 102 displayed a valgus deformity preoperatively, whereas 13 displayed varus deformity that required surgical repair; and these investigators found that there was a significant association between patient satisfaction and postoperative alignment (2). And, although nonunion is an often cited complication of triple arthrodesis, malposition is another potential factor that can compromise a successful long-term outcome. In an analysis of triple arthrodesis, misalignment of the fusion was noted to account for 66% of the failures (8). In another report of a series of 17 adults who underwent triple arthrodesis, 6 of the feet were noted to display moderate to severe valgus position of the heel following the surgery, and 4 of these developed ankle degenerative joint disease and 3 experienced persistent difficulty wearing shoes because of persistent valgus deformation of the hindfoot (6). The authors of the current report have, for several years, used a protocol for evaluating the alignment achieved by
the triple arthrodesis (14), and the purpose of the current investigation was to review subjective (pain) and objective (radiographic) outcomes in 24 patients (28 feet) who underwent a triple arthrodesis when these specific realignment considerations were used. It was interesting to note that triple arthrodesis led to what would be considered improvement in all of the radiographic variables, although only the changes in the AP talocalcaneal, AP talo-first metatarsal, and lateral talofirst metatarsal angles were statistically significant (Table 2). These observed changes attest to the power that the surgeon has in regard to altering the structural alignment of the foot when undertaking triple arthrodesis. Specifically, the data in the current study demonstrated a median change of 10.4 (range 1.0 to 26.2 ) for the lateral talo-first metatarsal angle, whereas a median change of 6.7 (range 8.0 to 23.0 ) was observed for the AP talo-first metatarsal angle (Table 3). It appears from these results that the average realignment of the forefoot to the hindfoot in the sagittal plane was greater than realignment in the transverse plane. These results are comparable to other reports of angular correction in the frontal and sagittal planes (2, 6). Moreover, the AP talocalcaneal angle displayed a median change of 6.86 (range 11.40 to 27.50 ), compared with a median 2.0 (range 34.8 to 20.4 ) change measured for the lateral talocalcaneal angle (Table 3); and these changes suggested that realignment within the hindfoot (between the talus and calcaneus) involved more transverse plane movement of the talus back over the calcaneus, rather than markedly changing the pitch of either the talus or the calcaneus in the sagittal plane. In fact, Table 3 also shows that the median changes in the calcaneal inclination and talar declination angels were 1.55 (range 9.30 to 6.70 ) and 5.00 (range 19.40 VOLUME 48, NUMBER 3, MAY/JUNE 2009
291
to 21.50 ), respectively, suggesting that what change did occur in the sagittal plane alignment of the hindfoot occurred primarily by means of reduction of the downward pitch of the talus rather than elevation of the pitch of the calcaneus. This observation is probably consistent with the general goal of establishing a stable, plantigrade foot by means of triple arthrodesis, and is probably more in keeping with reconstruction of the pes valgus deformity rather than repair of the cavus foot. No calcaneal osteotomies were required to achieve the desired amount of tarsal realignment, although we used autogenous bone graft at 6 (7.14%) arthrodesis sites and allograft at 1 (1.19%) calcaneocuboid site, and posterior soft tissue release in the form of Achilles tendon lengthening, or tenotomy, was performed in 21 (75.00%) of the procedures (Table 2). Again, we feel that these changes, which can be used to markedly alter the relationship of the forefoot to the hindfoot, and that of the hindfoot to the leg, are related primarily to the freedom that the surgeon has in regard to selecting the optimal orientation of the tarsals following resection of the midtarsal and subtalar joints. Nonunion rates have decreased to about 5% in the recent literature (2, 4, 7, 13), and we observed approximately the same prevalence of radiographic evidence of nonunion, namely 4 (4.76%) in 28 triple arthrodeses. Interestingly, 3 (60%) of the nonunions occurred at the calcaneocuboid joint, and we feel that this is probably the result of the obliquity of interfragmental compression screw fixation across the fusion interface, and the vulnerability of this fusion site to weightbearing forces sustained during the postoperative phase. In the future, it would be interesting to see if the addition of buttressing staples, or even a neutralization plate, might substantially decrease the prevalence of nonunion at this site (or the talonavicular fusion site, for that matter). The short-version McGill visual analog pain scale questionnaire was used to evaluate the preoperative and postoperative pain for 18 of the triple arthrodeses, because the patients failed to return the questionnaire following 10 of the procedures. Overall, we observed a statistically significant decrease in pain between the preoperative and postoperative periods, and 77.78% of the patients experienced none to mild pain long-term following the operation (Table 4). We also measured dependence between the radiographic variables and pain following triple arthrodesis using Spearman’s rank correlation coefficient, which is a distribution-free test of independence between 2 variables that makes no assumptions about continuity of the data, and assumes only that the data are ordered by rank. A look at Figures 6 and 7 shows that triple arthrodesis realigned the foot such that the likelihood of experiencing none to mild postoperative pain decreased if the preoperative lateral talocalcaneal angle was normal to begin with, and it increased if the preoperative lateral talar declination angle was normal. This first observation is somewhat intuitive, in that it makes sense to think that a more normal lateral talar declination angle would require less realignment in the sagittal plane, hence 292
THE JOURNAL OF FOOT & ANKLE SURGERY
the triple arthrodesis (or any reconstructive osteotomy or arthrodesis) would be more likely to satisfactorily establish a plantigrade foot, which would likely be less painful than one that remained in a pes valgus (or varus) malposition. Although not intuitively obvious, it was interesting to note that if the lateral talocalcaneal angle was normal to begin with, it is likely that it would be more difficult (inversely correlated) to alleviate pain following realignment via triple arthrodesis. The precise meaning of this finding remains obscure, however we suspect that it has to do with the interaction of corrective motion of the calcaneus relative to the talus in both the transverse and sagittal planes. Like all observational investigations, a number of methodological limitations probably threaten the validity of the conclusions that we have drawn from our results. The fact that not all of the patients returned the pain questionnaire following the operation meant that we observed a 35.71% loss to follow-up that, in all likelihood influenced our interpretation of the alteration of pain following with triple arthrodesis. Although we are not able to clearly state why the patients did not return the pain questionnaire, hence, lost to follow-up, we did not anecdotally identify any specific reason for the loss. Furthermore, in this investigation, the operating surgeons also measured and recorded the preoperative and postoperative radiographic values, and this could have imparted some degree of bias toward an effect of the surgery. In conclusion, triple arthrodesis appears to be a powerful method of realigning the forefoot to the hindfoot. It can be used to clinically and statistically significantly improve the structural alignment of a number of radiographic measurements that describe the orientation of the tarsals, namely the AP talocalcaneal angle, the AP talo-first metatarsal angle, and the lateral talo-first metatarsal angle. Our results also suggest that triple arthrodesis realigns the relationship of the talus and calcaneus primarily by means of transverse plane movement of the talus back over the calcaneus, rather than by altering the sagittal plane pitch of the calcaneus; and that change in the sagittal plane alignment of the hindfoot takes place primarily by means of reduction of the downward pitch of the talus rather than by means of elevation of the pitch of the calcaneus. Although we observed only 1 (1.19% of fusion sites, 3.57% of cases, 4.17% of patients) symptomatic nonunion in this case series, this complication remains a risk whenever arthrodesis is undertaken. In the future, the results of this case series may be useful in the development of randomized controlled trials or prospective cohort studies that further explore the risk factors and outcomes associated with triple arthrodesis.
Acknowledgment We appreciate the work of D. Scot Malay, DPM, MSCE, FACFAS, who performed the statistical analyses described in this investigation.
References 1. Beischer AD, Brodsky JW, Pollo FE, Peereboom J. Functional outcome and gait analysis after triple or double arthrodesis. Foot Ankle Int 20(9): 545–553, 1999. 2. Pell RF, Myerson MS, Schon LC. Clinical outcome after primary triple arthrodesis. J Bone Joint Surg Am 82-A:47–57, 2000. 3. Angus PD, Cowell HR. Triple arthrodesis: a critical long-term review. J Bone Joint Surg Br 68-B(2):260–265, 1986. 4. Saltzman CL, Fehrle MJ, Cooper RR, Spencer EC, Ponseti IV. Triple arthrodesis: twenty-five and forty-four-year average follow-up of the same patients. J Bone Joint Surg 81-A(10):1391–1402, 1999. 5. Sangeorzan BJ, Smith D, Veith R, Hansen ST. Triple arthrodesis using internal fixation in treatment of adult foot disorders. Clin Orthop 294: 299–307, 1993. 6. Graves SC, Mann RA, Graves KO. Triple arthrodesis in older adults: results after long-term follow-up. J Bone Joint Surg Am 75-A(3):355–362, 1993. 7. Fortin PT, Walling AK. Triple arthrodesis. Clin Orthop 365:91–99, 1999.
8. Maenpaa H, Lehto MU, Belt EA. What went wrong with in triple arthrodesis. Clin Orthop 391:218–223, 2001. 9. Bednarz PA, Monroe MT, Manoli A 2nd. Triple arthrodesis in adults using rigid internal fixation: an assessment of outcome. Foot Ankle Int 20(6):356–363, 1999. 10. Wenz W, Bruckner T, Akbar M. Complete tendon transfer and inverse Lambrinudi arthrodesis: preliminary results of a new technique for the treatment of paralytic pes calcaneus. Foot Ankle Int 29(7):683–689, 2008. 11. Malay DS. The PI Manual: A Handbook of Podiatric Medicine and Surgery, ed 2, pp 214 and 237, The Podiatry Institute, Decatur, GA, 2008. 12. Melzack R. The short-form McGill Pain Questionnaire. Pain 30: 191–197, 1987. 13. Wukich DK, Bowen JR. A long-term study of triple arthrodesis for correction of pes cavovarus in Charcot-Marie-Tooth disease. J Pediatr Orthop 9(4):433–437, 1989. 14. Catanzariti AR, Mendicino RW, Whitaker JM, Reeves CL. Realignment considerations in the triple arthrodesis. J Am Podiatr Med Assoc 95(1): 13–17, 2005.
VOLUME 48, NUMBER 3, MAY/JUNE 2009
293