Subjects
The study design was a retrospective review and was approved by our hospital's institutional review board. Between 2000 and 2004 at our institution, eighty-nine consecutive patients with cerebral palsy who had progressive midfoot break and excessive valgus deformity of the heel that was resistant to orthotic management underwent either a calcaneal lengthening osteotomy or an extra-articular arthrodesis of the subtalar joint. All operations were performed by a single surgeon (H.W.K.). Both procedures and the advantages and disadvantages of each operation were explained in detail before surgery. The choice of the procedure was based on the preference and informed consent of the patients and their parents. Most of the individuals who underwent extra-articular arthrodesis of the subtalar joint chose the procedure on the basis of their economic status; many parents wanted to have their children treated with extra-articular arthrodesis of the subtalar joint because autogenous cancellous bone graft was used and the cost of commercial allograft bone was not covered by the National Public Health System or their private health insurance company. Furthermore, many parents wanted to have their children treated with autograft rather than allograft (see surgical technique below).
Exclusion criteria for this study were an inability to walk without any assistance or orthosis prior to surgery; a history of previous foot surgery except an Achilles tendon lengthening; a history of additional and concurrent procedures on the forefoot and/or midfoot; a history of graft slippage in the early postoperative period, which necessitated subtalar fusion in addition to calcaneal lengthening osteotomy; a history of nonunion of the subtalar fusion as determined by nonvisualization of a solid osseous mass in radiographs made four months postoperatively; and a time interval between the operation and final follow-up of less than two years.
A total of eighty-one feet in forty-seven patients were included in the study. The patients had a mean age of eight years and one month (range, five years and six months to sixteen years and eight months) at the time of surgery. There were twenty-seven boys and twenty girls, and the mean duration of follow-up was thirty-nine months (range, twenty-six to sixty-one months). All patients had spastic diplegia and a crouch gait before surgery, defined by sagittal plane kinematic data showing ankle dorsiflexion of >15° and knee flexion of >30° during the late stance phase of gait25. Eleven patients (nineteen feet) had undergone a selective dorsal rhizotomy before the index operation. All patients had had pain on weight-bearing or they had a callosity under the plantar flexed talar head and resultant difficulty in wearing shoes and orthoses. The preoperative functional mobility as measured by the Gross Motor Function Classification System26 was level II in all patients. The index operation was performed as one of the multilevel and simultaneous surgical procedures on the lower extremities designed to improve the entire gait pattern, which was assessed by physical examination and three-dimensional gait analysis.
Subjects were divided into two groups. Group I consisted of thirty-seven feet in twenty-two patients who underwent calcaneal lengthening osteotomy. Group II comprised forty-four feet in twenty-five patients who underwent extra-articular arthrodesis of the subtalar joint. At the time of surgery, the patients had a mean age of seven years and eight months (range, five years and six months to thirteen years) in Group I and eight years and two months (range, five years and six months to sixteen years and eight months) in Group II. The mean postoperative follow-up period was forty-one months (range, twenty-six to fifty-seven months) in Group I and thirty-eight months (range, twenty-seven to sixty-one months) in Group II.
Surgical Techniques
The calcaneal lengthening osteotomy was performed as described by Mosca6. Briefly, the middle facet of the subtalar joint was identified with a Freer periosteal elevator by probing over the dorsum of the calcaneus. The instrument was then moved distally and was expected to slide into the interval between the anterior and middle facets that was the site of the planned osteotomy. With use of intraoperative fluoroscopy, a transverse saw cut was made through both cortices, and a laminar spreader was used to open the osteotomy and simultaneously restore coverage of the talonavicular joint. After the appropriate length and size of the graft to be inserted into the osteotomy site was determined, an autograft that had been obtained from the iliac crest in patients close to skeletal maturity (as assessed by closure of the triradiate cartilage) or a commercially available human iliac-crest allograft bone wedge was trimmed into a trapezoidal shape and matched in size to fit the osteotomy site. An autograft was used in seven feet, and an allograft was used in thirty feet. An appropriately sized Kirschner wire or Steinmann pin was placed longitudinally across the distal end of the calcaneocuboid joint to maintain the stability of the graft. Medial plication of the talonavicular joint capsule or the tibialis posterior tendon was not performed in any of the patients.
Extra-articular arthrodesis of the subtalar joint was performed with use of a modified Dennyson-Fulford technique17. Briefly, the calcaneus was rotated under the talus, and the hindfoot was held in the corrected neutral position with the ankle in equinus. A threaded guidewire was driven from the talar neck into the calcaneus in a posterolateral direction to emerge through its lateral cortex near the plantar surface. The reduction status of the talonavicular joint was confirmed under intraoperative fluoroscopic control. After drilling but not tapping the bone, a cannulated AO screw in an appropriate size (3.5, 4.5, or 7.0 mm in diameter depending on the size of the bone) was inserted to hold the hindfoot correction. Decortication of the nonarticular undersurface of the talus and upper surface of the calcaneus was done with use of a dental burr, after which the sinus tarsi was packed with cancellous bone chips.
A contracted peroneus brevis and gastrocnemius-soleus complex was best appreciated after correction of the hindfoot valgus deformity. Intramuscular lengthening or z-plasty of the contracted peroneus brevis tendon was performed in seventy-two feet. Twenty-one Vulpius or Strayer-type gastrocnemius recessions and fifty-two Achilles tendon lengthenings were performed in order to achieve approximately 10° of passive ankle dorsiflexion. Four patients had received a bilateral Achilles tendon lengthening at other institutions prior to the index operation. Postoperatively, all feet were immobilized in a long leg cast for four weeks, and then a short leg walking cast was used for another four weeks. For most patients, a posterior leaf-spring or floor-reaction ankle-foot orthosis was prescribed after the removal of the cast. All patients participated in a comprehensive rehabilitation program, with a gradual return to previous activities as tolerated.
Measurement of Outcome: Radiographs and Dynamic Foot-Pressure Measurements
For radiographic analyses, standardized weight-bearing anteroposterior and lateral radiographs of the feet and ankles21 made before surgery and at the final follow-up examination were used. Anteroposterior talus-first metatarsal and talocalcaneal angles were measured on the anteroposterior radiograph, while the lateral talocalcaneal angle, the calcaneal pitch angle, and the Meary angle (lateral talus-first metatarsal angle) were measured on the lateral radiograph, with use of the techniques described in the literature from which the normative data were generated27,28.
The F-Scan high-resolution pressure assessment system (Tekscan, South Boston, Massachusetts) was used to measure dynamic foot pressure. Pressure was recorded at 50 Hz with a pressure-sensitive insole consisting of a 0.15-mm-thick sensor with an embedded gridwork of 960 pressure-sensing cells distributed evenly at 0.5-cm intervals. Before its use, a disposable insole was trimmed to fit into the shoes. Patients walked approximately 20 m in order to acquaint themselves with the system. Foot pressures in nine areas, including the hallux, first metatarsal head, second metatarsal head, third and fourth metatarsal heads, fifth metatarsal head, medial aspect of the midfoot, lateral aspect of the midfoot, medial aspect of the calcaneus, and lateral aspect of the calcaneus, were recorded for five steps in the middle of the test walk, and the mean value was calculated. The pressure-reading data were saved and then were processed with custom-made software.
The pressure-time data for each individual area were graphed with use of a normalized pressure and time scale. The integrals of the pressure-time graphs indicate the total pressure achieved by each area of the foot, and the total pressure was identified in this study as impulse. The key foot (plantar) pressure parameter examined in this study was the relative vertical impulse, which is the impulse exerted on each area divided by the total impulses of all areas (Fig. 1). It denotes the portion of the work achieved in a particular area of the foot during gait, so the data would not be affected by body weight, plantar surface area, size of the foot, or contact time of the foot during walking20,21,23. Other parameters examined in this study included the anteroposterior index (the anteroposterior length of the path of the center of pressure divided by the anteroposterior length of the foot) and the center-of-pressure excursion index (the maximum lateral deviation of the center of pressure divided by the maximum width of the foot) for the evaluation of anteroposterior and lateral movements of the center of pressure23,24, respectively (Fig. 2). Also, in order to assess absorption of the impact caused by the pronation movement occurring in the subtalar joint, the presence or absence of medial deviation in the path of the center of pressure immediately after initial contact was observed22,23 (Fig. 2).
Statistical Analyses
In order to minimize measurement errors, two fellowship-trained pediatric orthopaedic surgeons (K.B.P. and S.Y.J.) performed all radiographic and pedobarographic measurements. All parameters were measured twice by each author, and they were then averaged. To determine the influence of the surgical procedure and to identify significant group differences, measured values from the radiographs and dynamic foot-pressure measurements made before surgery and at the latest follow-up evaluation were compared. The mixed-model method was used to allow for the repeated measurements from individual patients and to adjust for intraperson correlation to ensure the validity of the results. In addition, in order to adjust for the effects of age, sex, and time to assessment, we performed a multivariate analysis. Plantar pressure data obtained from patients were also compared with our normative database of fifty age-matched healthy volunteers. As to the existence of initial medial deviation in the path of the center of pressure (a binary response variable), the generalized estimating equation method was used. P values were adjusted for multiple comparisons with use of the Bonferroni method. The adjusted p value is the minimum value of (punadj × k) or 1, where punadj is the unadjusted value and k is the number of comparisons. The level of significance was set at p < 0.05.
At the final follow-up examination, all of our patients were found to have no pain or callosity under the head of the talus, and patient tolerance of the orthosis and shoes was improved. The valgus deformity of the heel had been corrected, and varying degrees of longitudinal arch created. For seven patients, the status according to the Gross Motor Function Classification System had improved by one level.
Radiographic Parameters
Before surgery, there were no significant differences between the groups in terms of radiographic parameters on the anteroposterior and lateral radiographs (Table I). In both groups, the anteroposterior talus-first metatarsal angle and the Meary angle were increased above normal, and the calcaneal pitch angle was lower than normal27. The anteroposterior talocalcaneal angle (p < 0.0001 in Group I and p = 0.0100 in Group II), the anteroposterior talus-first metatarsal angle (p < 0.0001 in Groups I and II), and the Meary angle (p < 0.0001 in Groups I and II) were improved after surgery. The calcaneal pitch angle was improved in Group I (p = 0.0053); however, it was not improved in Group II (p = 0.7531) (Fig. 3). The postoperative anteroposterior talocalcaneal and talus-first metatarsal angles were not significantly different between the groups; however, the postoperative lateral talocalcaneal and calcaneal pitch angles were higher in Group I and the Meary angle was lower in Group I (Table I).
Measurements of Dynamic Foot Pressure
Pressure-time curves recorded simultaneously under nine measurement areas demonstrated that two distinct peaks caused by heel strike and propulsion of the forefoot became evident in the stance phase after surgery (Fig. 4). Before surgery, no significant difference between the groups was detected with respect to the relative vertical impulse for each area of the sole, except the fifth metatarsal head (p = 0.0337) (Table II). In both groups, the relative vertical impulses for the hallux, first metatarsal head, and medial aspect of the midfoot were higher than the normal control, and the relative vertical impulses of the third and fourth metatarsal heads, fifth metatarsal head, lateral aspect of the midfoot, and calcaneus were lower than the normal control (Table II).
At the time of the final follow-up, the relative vertical impulses for the hallux (p = 0.0077 in Group I and p = 0.0228 in Group II), first metatarsal head (p = 0.0017 in Group I and p < 0.0001 in Group II), and medial aspect of the midfoot (p = 0.0007 in Group I and p < 0.0001 in Group II) in both groups were found to be decreased, while the relative vertical impulses for the lateral aspect of the midfoot (p = 0.0357 in Group I and p = 0.0009 in Group II), lateral aspect of the calcaneus (p < 0.0001 in Groups I and II), and medial aspect of the calcaneus (p < 0.0001 in Groups I and II) in both groups were found to be increased. A comparison between the groups revealed that postoperative relative vertical impulses for the first, second, and third and fourth metatarsal heads in Group II were lower than those both in Group I and in the normal controls (Table III). The relative vertical impulses for the medial aspect of the midfoot in Group II were higher than they were in Group I and in the normal subjects. Importantly, the relative vertical impulses for all areas, except the medial aspect of the midfoot, in Group I were similar to those of the normal controls (Table III).
With respect to the movement of the center of pressure before surgery, the anteroposterior index was not different between the two groups and the center-of-pressure excursion index was greater in Group II, although both the anteroposterior index and center-of-pressure excursion index were lower than normal (Table IV). At the final follow-up examination, the anteroposterior index in both groups was found to be increased (p = 0.0031 in Group I and p = 0.0035 in Group II); however, both were still lower than that of the normal controls. Postoperatively, the center-of-pressure excursion index in both groups was increased (p = 0.0167 in Group I and p = 0.0358 in Group II) and was not different from that of the normal controls. Initial medial deviation of the path of the center of pressure was noted in five feet (14%) in Group I and eight feet (18%) in Group II before surgery. At the time of the latest follow-up, twenty-six feet (70%) in Group I and twenty-one feet (48%) in Group II were found to have medial deviation of the path of the center of pressure. Compared with the values before surgery, the odds that a foot would have a medial deviation of the path after surgery were 62.5 times greater in Group I (p < 0.0001) and 10.1 times greater in Group II (p = 0.0035).
When a tight gastrocnemius-soleus complex pulls on the hindfoot, the foot will break through its midportion in an attempt to accommodate the plantar flexion and still maintain the foot flat on the floor during the stance phase of gait. In this setting, normal dorsiflexion of the ankle is prevented and the dorsiflexion stress is shifted to the talonavicular joint. The soft tissues underlying the head of the plantar flexed talus are subjected to excessive direct axial loading and shear forces, allowing posterior movement of the calcaneus on the talus and more external rotation and dorsiflexion of the foot1,3,6. Kinematic gait data may show increased ankle dorsiflexion during stance; however, this is due to midfoot breakage, and the contracted gastrocnemius-soleus complex becomes more evident after correction of a foot deformity.
The objective of a calcaneal lengthening osteotomy is to lengthen the calcaneus, dorsiflex the talus, invert the subtalar joint, and align the midtarsal joints to provide a rigid lever. It has been known to align the talonavicular joint and correct external rotation and lever arm dysfunction through the subtalar joint. Although the long-term outcomes of this procedure are still pending, there is no doubt that a calcaneal lengthening osteotomy has the added bonus of maintaining joint motion2,3,6,11,12. In an extra-articular arthrodesis of the subtalar joint, bone grafts are placed into the sinus tarsi with the heel in a corrected position. The presumptive disadvantage of an extra-articular arthrodesis of the subtalar joint is the development of arthritis in the adjacent joints. However, Bourelle et al. reported the long-term results of the original Grice-Green procedure in patients with cerebral palsy, and noted no deformity of the talus or arthritis of adjacent joints at skeletal maturity5. Comparative studies on the differences in the operative results between these two procedures are still lacking.
Our results show that both procedures correct abduction of the forefoot and subluxation of the talonavicular joint, as represented by postoperative reduction of the anteroposterior and lateral talus-first metatarsal angles. The reduction of the anteroposterior talocalcaneal angle suggested the correction of hindfoot valgus deformity, and a decreased Meary angle implied improvement of malalignment of the talonavicular joint and an elevated medial longitudinal arch. Postoperative correction of flexion of the talus can be interpreted as a reduced lateral talus-first metatarsal angle9,27. In the present study, plantar flexion of the talus was corrected after an extra-articular arthrodesis of the subtalar joint; however, the calcaneal pitch was not similarly improved. In a valgus hindfoot deformity, both the talus and calcaneus are in equinus, but the talus is in relatively greater equinus. We believe that simply holding the heel in a corrected neutral position and the ankle in equinus is not sufficient for three-dimensional correction of the deformity.
Calcaneal pitch was significantly improved following calcaneal lengthening osteotomy. Although the mechanism by which calcaneal lengthening corrects valgus deformity of the hindfoot and its relationship to the restoration of the medial longitudinal arch are not clearly known, increased tension in the plantar ligaments and increased length of the plantar fascia have been suggested to be key elements for creating the bowstring effect that restores the arch6,29,30. The distal aspect of the calcaneus, along with the cuboid, navicular, and spring ligament, is pushed distally, and the posterior part of the calcaneus is pushed posteriorly. The central plantar fascia, being of a fixed length, resists straight distal distraction at the site of the osteotomy. The arch is then elevated, and the subtalar joint complex is inverted in a manner suggestive of the Jack toe-raising test6,30,31.
Dynamic foot-pressure measurement has been regarded as a quantitative measure capable of universally describing the function of the foot, and it represents the overall ability of patients to support and efficiently transfer their body mass during walking20-24. With both procedures studied, pressure-time curves recorded simultaneously under nine measurement points demonstrated postoperative improvement of foot pressure-point progression in time sequence from heel to forefoot, suggesting more efficient transfer of body mass during walking.
The two main functions of the subtalar joint are to absorb forces generated by the foot striking the ground and to maintain balance while walking on uneven surfaces19. However, clinical measurement of the mobility of the subtalar joint is neither accurate nor reproducible18, and it has been found to be not associated with radiographic findings31,32. In reality, it is not reasonable to average foot-pressure data in a manner comparable with radiographic measurements because of the wide variation in the shapes and sizes of the feet documented by dynamic foot-pressure analysis. Furthermore, it has been shown that the differences in body height and mass influence gait velocity and plantar pressure distribution11,12. In the present study, we calculated the relative vertical impulse in order to describe the relative load-distribution pattern that is independent of anthropometric and spatiotemporal parameters20,21. We observed no differences between Groups I and II in the preoperative relative vertical impulse in each area, except for the fifth metatarsal head, and we found no differences in the radiographic measurements between the groups, indicating that the plantar pressure distribution was similar in both groups and that the feet in both groups were similarly deformed. The relative vertical impulses of the hallux, first metatarsal head, and medial aspect of the midfoot were decreased, whereas those of the lateral aspect of the midfoot and calcaneus were increased, after both calcaneal lengthening osteotomy and extra-articular arthrodesis of the subtalar joint, indicating postoperative elevation of the medial longitudinal arch.
Ledoux and Hillstrom determined the distribution of plantar vertical forces in a neutrally aligned normal foot and a flexible flatfoot and observed that flatfeet had more force in the subhallucal area24. They attributed this finding to inappropriate action of the peroneus longus due to malalignment of the foot and abnormal activation of the flexor hallucis longus as compensation. This might be indicative of the aberrant first ray mechanics in flatfoot33. Postoperative normalization of the relative vertical impulse of the hallux in both of our groups suggests the disappearance of abnormal activation of the flexor hallucis longus after surgery.
There is a secondary supination deformity of the forefoot in relation to the hindfoot in longstanding planovalgus foot deformity in children with cerebral palsy who are able to walk. Uncorrected forefoot supination can result in a postoperative lack of weight-bearing on the medial aspect of the forefoot, thereby causing recurrence of the valgus deformity of the hindfoot over time because of loss of the tripod effect of the foot6. A lower than normal postoperative relative vertical impulse for the first metatarsal head in Group II can be interpreted as a consequence of uncorrected forefoot supination after the extra-articular arthrodesis of the subtalar joint. On the contrary, the relative vertical impulse for the first metatarsal head was also significantly decreased after a calcaneal lengthening osteotomy in the present study, and it became essentially equal to normal controls. These findings suggest that the calcaneal lengthening osteotomy corrects the forefoot supination deformity, which might be attributed in part to restored function of the peroneus longus, a powerful plantar flexor of the first metatarsal. The lateral column of the planovalgus foot is short relative to the medial column, and, as a consequence, the lever arm for the peroneus longus becomes short. With a calcaneal lengthening osteotomy, the lever arm of the peroneus longus is lengthened, and, therefore, the peroneus longus is effectively shortened, correcting the forefoot deformity by means of a tenodesis effect6,11,33.
A calcaneal lengthening osteotomy is not a treatment for a planovalgus foot deformity, but rather a treatment for valgus deformity of the hindfoot6. The objective of a calcaneal lengthening osteotomy is to rotate the foot with respect to the talonavicular axis in both the horizontal and sagittal planes, thus inducing combined adduction and plantar flexion of the foot at the talonavicular joint. Although forefoot supination is best appreciated after correction of the hindfoot, additional procedures to fully correct the residual deformity in the forefoot or the medial column of the foot may depend on the degree of deformity, the functional demands of patients, or the surgeon's judgment. Further studies are necessary to answer the questions of when, how, and to what extent medial column reconstruction should be combined with calcaneal lengthening osteotomy.
The path of the center of pressure is a virtual graphic line indicating the instantaneous point of application of the resultant vertical foot-floor force during the gait cycle. Normally, it initiates at the calcaneus, deviates to the lateral side of the foot, and terminates near the second metatarsal head23,24. Our finding of an increased center-of-pressure excursion index after both procedures is indicative of a shift of the path of the center of pressure through the lateral aspect of the midfoot after initial contact and elevation of the medial longitudinal arch. Also, an increased anteroposterior index for the path of the center of pressure after both surgical procedures implies an improved ankle plantar flexion and knee extension couple, resulting from release of a tight gastrocnemius-soleus musculotendinous unit.
A short and slight deviation of the path of the center of pressure toward the medial side immediately after heel strike implies shock absorption and dissipation of the forces generated by the foot striking ground19,23,34. As expected, pronation movement of the subtalar joint in the present study was more distinct in the feet treated with a calcaneal lengthening osteotomy. Interestingly, initial medial deviation of the path of the center of pressure was observed in 48% of the feet treated with an extra-articular arthrodesis of the subtalar joint, suggesting the possibility that there might be shock absorption at the subtalar joint even in the presence of an extra-articular fusion of the joint or that there might be compensatory movement of the ankle joint in the coronal plane.
We noted several limitations to our study. Its retrospective design did not allow direct comparison of the technique of calcaneal lengthening osteotomy and extra-articular arthrodesis of the subtalar joint with regard to function or health-related quality of life. Since the Gross Motor Function Classification System levels before surgery were the same in all patients, we cannot suggest which procedure is preferable in patients with different functional levels. Furthermore, as postoperative gait analysis was performed in only nineteen patients, we cannot report any dynamic parameters of the hip and knee joints that might have influenced the results.
From the present study, an extra-articular subtalar arthrodesis appears to be an effective means to achieve predictable correction of heel valgus; however, supination deformity of the forefoot remained and calcaneal equinus is not corrected. On the contrary, we consider a calcaneal lengthening osteotomy to be the treatment of choice because postoperative foot-pressure distribution was found to more closely approximate the normal pressure distribution. However, additional long-term randomized trials of the calcaneal osteotomy with and without reconstruction of the medial column of the foot are warranted in order to answer the question of longevity of the procedure. In addition, investigation of the association between improvements in the foot-pressure distribution and changes in the function of the foot is needed. 