The study design was approved by our institutional review board. Inclusion criteria for the study were (1) tibial lengthening by distraction osteogenesis with the use of an Ilizarov fixator; (2) fixation of the fibula to the frame both proximal and distal to the corticotomy at least during tibial lengthening; (3) the absence of preoperative angular deformities in the foot and ankle; (4) the absence of complications that might affect fibular length, such as cutting out or breakage of a tibiofibular transfixing wire during distraction, an incomplete fibular corticotomy, or premature fibular consolidation; and (5) limb length inequality of <1 cm both before and after bilateral tibial lengthening.
From 1995 to 2004, seventy-four tibiae in thirty-seven patients who underwent bilateral tibial lengthening for short stature at our institution met the inclusion criteria. The mean age at the time of surgery was 21.7 years (range, thirteen to thirty-one years). Nineteen patients were male and eighteen were female. The diagnosis was idiopathic short stature in twenty-three patients, achondroplasia in six, hypochondroplasia in two, Turner syndrome in two, and metaphyseal chondrodysplasia, dyschondrosteosis, hypothyroidism, and hypopituitarism in one each (Table I). The average duration of follow-up after the completion of the lengthening procedure (complete consolidation of the regenerate) was forty-five months (range, thirty to sixty-two months).
Surgical Techniques
All procedures were performed by one surgeon. A preconstructed Ilizarov circular frame was applied with tensioned wires and Schanz screws. In all patients, the fibula was fixed with at least one tibiofibular transfixing wire passing through the fibular head proximally and another passing through the lateral malleolus distally at the most proximal and distal ring, respectively. Fibular osteotomy was performed at the distal third with use of a multiple drill-hole technique without segmental resection. An approximately 2-cm-long skin incision was made on the lateral aspect of the fibula, and then the lateral side of the periosteum was elevated carefully. The cortex was predrilled in multiple directions with a small-diameter drill. Finally, a 5-mm osteotome was advanced to complete the osteotomy. The elevated periosteum was repaired with absorbable sutures after the completion of the osteotomy.
On the basis of the surgical procedure performed on the tibia, the legs were categorized into one of three groups. Group I included eleven legs in which a monofocal tibial corticotomy was performed just below the tibial tuberosity (Fig. 1). Group II consisted of thirty-nine legs in which the tibia was lengthened with use of bifocal corticotomies, distal to the tibial tuberosity and at the supramalleolar level, and the fibula was lengthened by monofocal osteotomy in its distal third. In the eight legs that had tibia vara, a concomitant angular correction was performed at the proximal corticotomy site (Fig. 2). Group III comprised twenty-four legs that had tibial lengthening over an intramedullary nail, with a monofocal corticotomy performed at the proximal metaphyseal-diaphyseal junction of the tibia6.
Distraction commenced between the seventh and the tenth postoperative day at a rate of 1 mm per day in Groups I and III and at 1 mm per day at both the proximal and distal corticotomy sites in Group II. The distraction rate was decreased if there was radiographic evidence of poor bone formation or if the patient had excessive pain or neurologic symptoms during lengthening. In Groups I and II, the Ilizarov fixator was removed when more than three cortical regeneration bridges across the tibial corticotomy site were confirmed on the radiographs, and a long leg cast was applied for four weeks. In Group III, after the desired length had been achieved, two distal interlocking screws and one distal tibiofibular transfixing screw were inserted and the Ilizarov fixator was removed. The distal tibiofibular transfixing screw was removed after the consolidation of the fibula.
Assessment of Ankle Valgus Deformity and Proximal Migration of the Fibula
To evaluate proximal migration of the fibula, the malleolar tip distance was measured in millimeters. The malleolar tip distance is the vertical distance between the tip of the lateral malleolus and the tip of the medial malleolus parallel to the mid-diaphyseal line of the distal aspect of the tibia1,7. The amount of proximal migration was assessed by comparing the malleolar tip distances before the surgery and at the time of the latest follow-up.
To evaluate ankle valgus deformity, the tibiotalar angle was measured at the intersection of the mid-diaphyseal line of the tibia and a line drawn across the flat subchondral line of the talar dome8. In this study, we defined ankle valgus deformity as a valgus change of =5° in the tibiotalar angle after tibial lengthening.
The malleolar tip distance and the tibiotalar angle were measured on the standing anteroposterior radiographs of the leg. The protocol for anteroposterior standing radiographs at our hospital requires that the patient stand in a bipedal stance with the patella facing forward in front of the film cassette, so that the calf touches the cassette and the radiographic tube is positioned 305 cm away and centered on the middle of the legs. In this retrospective study, serial changes in the malleolar tip distance during distraction were not assessed because radiographs during the distraction period were made with the Ilizarov ring fixator in situ, which inevitably placed some distance between the leg and the film cassette, affecting the magnification.
Data Analysis
Since data were collected from both limbs of each patient, comparisons were made with use of a linear mixed model and a marginal model with a generalized estimating equation to allow for the repeated measurements from individual patients and to account for intrapatient correlation.
To determine whether proximal migration of the distal end of the fibula caused ankle valgus deformity, we evaluated the relationship between the development of ankle valgus deformity and changes in the malleolar tip distance after the lengthening procedure, using a linear mixed model.
To assess how much migration was critical for the development of secondary ankle valgus deformity, we generated a receiver operating characteristic curve with changes in the malleolar tip distance as a test variable and development of ankle valgus deformity as a static variable. The optimal cutoff value, at which changes in the malleolar tip distance best differentiated between legs with and without ankle valgus deformity, was determined by estimating the minimal value of (1 — sensitivity)2 + (1 — specificity) and the maximum Youden index (J = sensitivity + specificity — 1) on the receiver operating characteristic curve9.
To determine what factors were associated with fibular migration over the critical amount in an Ilizarov tibial lengthening, we tested patient age at the time of surgery, sex, underlying diagnosis, length gain in centimeters and percentage, lengthening index, distraction rate, number of tibial corticotomies, the addition of a tibiofibular transfixing screw, and the development of fibular nonunion in relation to the changes in the malleolar tip distance of more than a critical amount as an event of interest. A marginal model with a generalized estimating equation was used with adjustment for age, sex, and underlying diseases. The level of significance was set at p < 0.05.
Source of Funding
No external sources of funding or support were utilized in the production of this work.
The mean tibial length gain was 6.9 cm (range, 4.7 to 11.5 cm), which represented a mean increase of 21.7% (range, 13.7% to 32.7%). The average rate of distraction was 1.24 mm per day (range, 0.75 to 1.78 mm per day), and the average lengthening index was 1.5 mo/cm (range, 0.8 to 2.2 mo/cm) (Table I). In all eight tibiae with a proximal varus deformity of an average 16.5° (range, 14° to 19°), concomitant angular correction at the proximal corticotomy site was achieved uneventfully.
Fibular nonunion developed in ten (14%) of seventy-four legs. All of the patients with a fibular nonunion had tenderness and motion at the site of nonunion after removal of the cast; however, these symptoms resolved with time. Three of the ten patients continued to have vague pain at the site of nonunion after running or other vigorous physical activities. However, the symptom was tolerable; therefore, no surgical intervention had been performed to address fibular nonunion at the time of the latest follow-up.
Ankle valgus deformity developed in six (8%) of the seventy-four legs, with an average change in the tibiotalar angle of 7.8° (range, 5° to 12°). Five of the six legs also had a fibular nonunion, and one had ankle valgus deformity develop without fibular nonunion. The patients with ankle valgus deformity had vague pain on both the medial and lateral sides of the ankle and valgus instability of the ankle especially when running and climbing the stairs. Surgical intervention to realign the lateral malleolus was recommended; however, all six patients declined the additional surgery at the time of the latest follow-up.
In addition to fibular nonunion and ankle valgus deformity, ninety complications developed during lengthening procedures. Nineteen pin-track infections were resolved with daily dressings and intermittent oral antibiotics. Forty-nine transfixing wires broke, and thirty-six of them were exchanged so that effective distraction was not jeopardized. In six legs with a knee flexion contracture, range of motion was fully regained through physiotherapy. Among ten ankle plantar-flexion contractures, five resolved with physical therapy, four were treated with an Achilles tendon lengthening, and one remained untreated. Six transient peroneal nerve palsies developed during the lengthening procedure, and all resolved within a few days after the distraction rate was decreased.
Among the three groups, the percentage of length gained was greater in Group II than in the two other groups (p = 0.001). Also, the lengthening index was lower (p < 0.001) and the distraction rate was faster (p < 0.001) in Group II with bifocal tibial lengthening (Table I).
Relationship Between Changes in the Malleolar Tip Distance and Ankle Valgus Deformity
Changes in the malleolar tip distance after tibial lengthening were <1 mm in twelve legs, 1 to <5 mm in forty-four, and =5 mm in eighteen. The malleolar tip distance at the latest follow-up evaluation was >10 mm in twelve legs, 7 to 10 mm in thirty-three, and =6 mm in twenty-nine. All ankle valgus deformities developed when the malleolar tip distance measured at the latest follow-up evaluation was =6 mm and the change in the malleolar tip distance was =5 mm (Figs. 3 and 4). The average change in the malleolar tip distance was 2.9 ± 2.7 mm for all legs and 6.3 ± 2.3 mm for the six legs with ankle valgus deformity (Table II). Linear mixed model analysis (p = 0.004) and the area under the receiver operating characteristic curve (0.890; 95% confidential interval, 0.812 to 0.968) revealed a significant relationship between change in the malleolar tip distance and development of ankle valgus deformity (Table II).
Determination of Optimal Cutoff Point of Changes in the Malleolar Tip Distance
The minimal value of (1 - sensitivity)2 + (1 - specificity)2 and the maximal Youden index on the receiver operating characteristic curve were estimated as 0.03 and 0.82, respectively, and the corresponding cutoff point was 5 mm (Table III). Therefore, we set 5 mm as an optimal cutoff value for changes in the malleolar tip distance that best differentiated between legs with and those without ankle valgus deformity after tibial lengthening.
Factors Associated with Changes in the Malleolar Tip Distance of =5 Millimeters
The number of tibial corticotomies (p = 0.043), the distraction rate (p = 0.034), and the development of a fibular nonunion (p = 0.039) were significantly different between the legs with a change in the malleolar tip distance of =5 mm and the legs with a change of <5 mm (Table IV). Bifocal tibial corticotomy, a rapid distraction rate, and the development of a fibular nonunion were regarded as factors associated with changes in the malleolar tip distance of =5 mm.
Note: The authors thank Dr. Hae-Ryoung Song, Biostatistician, Department of Biostatistics, Yonsei University College of Medicine, for the statistical analysis.