Ninety-two patients were managed with the Ilizarov technique at our institution between August 1988 and July 1994. Seventy patients were excluded from the present study because they either had a congenital deformity or did not have lengthening of the extremity. The remaining twenty-two patients (twenty-five limb segments) had limb-lengthening with or without simultaneous correction of deformity. The limb-length discrepancy was secondary to a pathological bone disorder in eight patients (nine limb segments) (Group 1), to post-traumatic growth arrest in seven patients (nine limb segments) (Group 2), and to growth arrest following a bone infection in seven patients (seven limb segments) (Group 3).
The demographic data, diagnosis, associated osseous disorders, clinical course, duration of lengthening, type of fixation used, duration of follow-up, and complications were determined from the medical records. Scanograms and radiographs were reviewed to measure the extent of the limb-length discrepancy and the results of lengthening. During the lengthening period, radiographs were made every four weeks until there was evidence of cortical bone formation on three cortices. If the residual discrepancy was more than one centimeter, another set of scanograms was made. Radiographs were also made three, six, and twelve months after removal of the fixation device.
Of the eight patients in Group 1, two had enchondromatosis and one each had Maffucci syndrome, melorheostosis, osteogenesis imperfecta, hypophosphatemic rickets, monostotic fibrous dysplasia, and irradiated bone (Table I). Seven tibiae, one femur, and one humerus were lengthened. Of the nine segments, four (in Cases 2, 6, and 7) were corrected at two levels and five were corrected at one. Six segments (in Cases 2 through 6) had simultaneous correction of an angular or rotational deformity. One patient (Case 2), who had enchondromatosis, had two separate procedures: a valgus deformity of the tibia and seven centimeters of shortening were corrected when the patient was six years old, and a residual deformity and a remaining eight centimeters of shortening were corrected when she was twelve years old. Another patient (Case 8) had 0.3 centimeter of lengthening of the tibia with an Ilizarov apparatus at another hospital and was subsequently transferred to our institution because of a major pin-site infection that required removal of the apparatus, débridement, and reapplication of the fixator.
Of the seven patients in Group 2, two (Cases 10 and 14) had two separate lengthening procedures (Table I), and two (Cases 14 and 15) had simultaneous correction of a deformity.
Of the seven patients in Group 3, two (Cases 20 and 22) had simultaneous correction of an angular deformity and one (Case 22) had correction at two levels. Three patients (Cases 16, 17, and 18) planned to have a second lengthening when they were older.
The Ilizarov device was used for twenty-two segments, the Wagner apparatus was used for two segments (in Cases 1 and 10), and an Orthofix fixator was used for one segment (in Case 9). The type of apparatus was chosen according to the preference of the surgeon. Distraction was initiated three to five days postoperatively, and the rate of lengthening was altered on the basis of the extent of bone formation. Once lengthening was complete, the fixation was secured in a neutral position until the osteotomy site was seen to be consolidated on radiographs. The limb was then placed in an above-the-knee cast or a knee-ankle-foot orthosis. The duration of treatment with the cast or brace depended on the preference of the surgeon but was usually continued until the osseous architecture was completely restored (normal osseous cortices and medullary cavity).
We defined the complications on the basis of the recommendations of Velazquez et al. Any unwanted event reported by a patient was considered as a complication regardless of whether any hospitalization or intervention was necessary. The complications were classified as minor when they responded to non-operative treatment and had no known lasting sequelae, and they were classified as major when they necessitated an additional operative procedure, prolonged treatment, or affected the outcome. Complications were categorized as osseous, soft-tissue, neurovascular, or psychological18.
Group 1
The mean age of the four boys and four girls was twelve years (range, six to sixteen years) (Table I). The mean duration of lengthening was six months (range, four to nine months) per limb segment. The mean duration of follow-up was four years and four months (range, one year and three months to six years and three months).
Preoperatively, the limb-length discrepancy averaged 6.0 centimeters (range, 2.7 to 8.8 centimeters). A mean of 6.2 centimeters (range, 2.7 to 9.0 centimeters) of lengthening was achieved, for a mean increase of 23 per cent (range, 9 to 35 per cent) of the initial length. The 0.3 centimeter of lengthening that was the result of an operation performed in another hospital (Case 8) was not included in the present assessment. Of the nine segments, seven (in Cases 1, 2, 3, 6, 7, and 8) had a final length discrepancy of 0.5 centimeter or less. The patient (Case 2) who had lengthening in two stages had an initial gain of seven centimeters and a later gain of eight centimeters. Equal limb lengths were achieved initially in one patient (Case 4), who had melorheostosis; however, twenty-eight months later, vascular complications and a persistent flexion contracture necessitated amputation through the knee. The mean lengthening index (total duration of treatment in months per centimeter of lengthening) was 1.1 (range, 0.6 to 2.5).
Complications
There were forty-one complications (mean, five complications; range, two to seven complications) in the nine segments (Table I). There were twenty-five minor complications (mean, three complications; range, two to four complications) and sixteen major complications (mean, two complications; range, zero to five complications). Only one of the nine segments (in Case 2) was not associated with a major complication. The sixteen major complications in the eight segments led to four additional operative procedures, and at the time of writing two additional procedures were being considered.
Osseous complications: Five patients (five segments) had an osseous complication. Two patients (Cases 3 and 6) had delayed consolidation at the site of the osteotomy. In three patients (Cases 1, 2, and 5), the treated bone fractured when the device was removed. The fracture healed in two patients (Cases 1 and 2) with immobilization in a cast. A non-union developed in the remaining patient, despite internal fixation with a Rush rod, and bone-grafting was performed; the fracture ultimately healed, leaving a residual limb-length discrepancy of 1.2 centimeters.
Soft-tissue complications: Seven patients (eight segments) had at least one soft-tissue complication. Four patients (Cases 1, 2, 7, and 8) (five segments) had a superficial pin-site infection, which was treated with oral or intravenous administration of antibiotics. There were no cases of cellulitis or osteomyelitis. Four patients (Cases 2, 4, 5, and 7) (four segments) had a flexion contracture of the knee. Of these patients, three (Cases 2, 5, and 7) had no residual contracture after cessation of the lengthening process and one (Case 4) had an amputation through the knee because of associated vascular problems. Stiffness of the shoulder and elbow developed in one patient (Case 3) during lengthening of the humerus and resolved with physiotherapy. Four patients (four segments) had a residual angulation deformity of the lengthened extremity after the fixation device was removed: the fixator was reapplied to correct the deformity in one patient (Case 2), one patient (Case 4) had an amputation, and two patients (Cases 7 and 8) were possible candidates for operative intervention in the future. Peripheral edema developed during fixation in five segments in five patients (Cases 1, 3, 4, 5, and 8); it resolved after removal of the fixation device.
Neurovascular complications: Neurovascular problems occurred in three patients (three segments). Two patients (Cases 6 and 7) had dysesthesias in the distribution of the common peroneal division of the sciatic nerve. In another patient (Case 4), pressure sores in the popliteal fossa and foot as well as a transient peroneal nerve palsy developed during the lengthening. The nerve palsy resolved after adjustment of the apparatus, and the pressure sores initially improved after adjustment of the apparatus and removal of the foot compartment of the frame. Additional pressure sores developed while the patient was wearing the above-the-knee brace, and an amputation through the knee was ultimately required because of vascular complications.
Psychological complications: None of the patients had overt psychological problems that necessitated premature removal of the fixation device. All patients complained of pain at the pin sites. One patient (Case 1) had problems with pain control due to a persistent pin-site infection; this was considered a major complication. The patient was admitted to the hospital and was managed with intravenous administration of antibiotics, oral administration of analgesics, and removal of the pin.
Group 2
The four boys and three girls were a mean of thirteen years old (range, nine to seventeen years old) (Table I). The mean duration of treatment was seven months (range, one to ten months) per limb segment, and the mean duration of follow-up was three years and four months (range, one year and one month to eight years). The mean preoperative limb-length discrepancy was 5.0 centimeters (range, 3.0 to 9.0 centimeters). A mean of 5.0 centimeters (range, 3.0 to 8.3 centimeters) of lengthening was achieved. Of the nine segments, six (in Cases 9, 10, 12, 13, and 15) had a final length discrepancy of 0.5 centimeter or less. Two patients (Cases 10 and 14) had two separate lengthening procedures. The mean lengthening index was 1.4 months per centimeter (range, 0.2 to 2.7 months per centimeter).
Complications
There were twenty minor and six major complications (mean, three complications [two minor and one major] per segment) (Table I). Two patients (two segments) had an osseous complication: one (Case 9) had a malunion, and one (Case 11) had a residual limb-length discrepancy of 1.2 centimeters. All seven patients had at least one soft-tissue complication, including a pin-site infection (eight segments), stiffness of the knee (three segments), a wound infection (one segment), and edema (one segment). A peroneal nerve palsy developed in one patient (Case 9) and resolved within a few days after the rate of distraction was decreased temporarily. All patients had pain postoperatively; two (Cases 12 and 15) had problems with pain control and needed prolonged hospitalization. Three segments in two patients (Cases 13 and 14) were not associated with a major complication.
Group 3
The three boys and four girls had a mean age of ten years (range, three to seventeen years) (Table I). The mean duration of treatment was eight months (range, four to thirteen months) per limb segment, and the mean duration of follow-up was three years and eleven months (range, one year and two months to six years and one month).
Preoperatively, the limb-length discrepancy averaged 10.1 centimeters (range, 3.4 to 20.0 centimeters). A mean of 9.9 centimeters (range, 3.0 to 19.5 centimeters) of lengthening was achieved. Of the seven segments, five (in Cases 16, 18, 19, 20, and 22) had a final length discrepancy of 0.5 centimeter or less; however, at the time of the most recent follow-up, it was thought that three segments (in Cases 16, 17, and 18) would most likely need additional lengthening to correct a residual discrepancy at skeletal maturity. The mean lengthening index was 1.0 month per centimeter (range, 0.3 to 2.0 months per centimeter).
Complications
There were twenty-two complications (fourteen minor and eight major), and a mean of three complications (two minor and one major) per segment. Two patients (Cases 21 and 22) had a fracture after removal of the fixation device, and a malunion developed in one (Case 21). All seven patients had at least one soft-tissue complication, including a pin-site infection (five segments), joint contracture (two segments), and stiffness of two joints (one segment). No patient had edema. Neurovascular complications included postoperative disseminated intravascular coagulation in one patient (Case 22), which was treated successfully. All seven patients had pain postoperatively; one patient (Case 20) had problems with pain control and needed prolonged hospitalization. Four segments (in Cases 16 through 19) were not associated with a major complication.
Several methods of limb-lengthening have been developed since the first lengthening procedure by Codivilla was documented in the English-language literature in 1905; however, these techniques have been associated with unacceptably high rates of complications6. The Wagner method appeared to solve many of the technical problems associated with limb-lengthening; however, it also has been associated with a very high rate of complications14. Currently, the most frequently used technique for limb-lengthening is based on Ilizarov's principles of distraction osteogenesis9-11. We believe that this technique is superior to other methods since bone-grafting and internal fixation are not required and associated deformities and malalignment of the extremity can be corrected without additional operations13. The rate of complications also appears to be lower than that for previously used methods5,6,15,19.
In the present study, the final length discrepancy was 0.5 centimeter or less in seven of the nine limb segments (eight patients) affected by an underlying bone disorder (Group 1). Bonnard et al. reported restoration of length to within five millimeters in twenty-three (88 per cent) of twenty-six segments that had normal bone and a mean overall increase in length of 15 per cent. Our results suggest that the percentage of limb equalization obtained in these eight patients was acceptable, taking into account the underlying abnormality of the bone, the high percentage (mean, 23 per cent) of length obtained, and the difficulty in estimating the final limb-length discrepancy in growing children who have dysplastic bone. Our results indicate that limb length was restored to within 0.5 centimeter of that on the contralateral side more effectively in the patients who had a bone disorder than in the two control groups (Groups 2 and 3) (Table II). These findings are difficult to interpret since additional lengthening was planned for three patients in Group 3 at the time of writing.
On the basis of the classification described by Velazquez et al., we reported a mean rate of five complications (three minor and two major) in the nine segments in Group 1. Velazquez et al. reported a much lower mean rate of 1.5 complications (0.9 minor and 0.6 major) in forty patients who had normal bone. In the present study, the rate of complications was higher in the patients who had an underlying bone disorder than in the control groups (Table II). The difference between these groups was in the nature and number of the major complications.
Thus, although the number of patients in our series was too small for statistical analysis, our findings demonstrated a higher rate of complications associated with lengthening of abnormal bone. Because the rates of complications associated with lengthening have been reported to be quite high5,6,15,16,18,19, it remains difficult to assess accurately the safety of this technique in patients who have a bone disorder. For these reasons, the rates of complications reported in the present study should not be interpreted as suggesting that it is substantially safer to perform lengthening through histologically normal bone. Rather, they may illustrate an important trend. Larger, prospective, long-term studies are needed to assess properly any statistical difference in the rates of complications between these two groups.
As stated, in the present study the patients who had an underlying bone disorder had more major complications than the other two groups. The most devastating complication occurred in the patient who had melorheostosis (Case 4); neurovascular and soft-tissue complications necessitated an amputation through the knee. Although Marshall and Bradish reported an uncomplicated lengthening in a patient who had melorheostosis, Younge et al. described problems with distal ischemia. Our results emphasize the need for caution in managing patients who have this condition.
There was also a trend for a higher prevalence of soft-tissue complications in the patients who had a bone disorder compared with those in the control groups. We believe that this trend may be partly explained by the skin and soft-tissue problems associated with these conditions rather than by the nature of the osseous lesion itself.
It is difficult to predict the growth remaining in a bone that has been lengthened and the effect of lengthening on the growth plate. Moreover, the effects of lengthening on future bone growth are more unpredictable the further a patient is from skeletal maturity. Additional research is necessary to establish a relationship between the age of the patient and the need for repeated lengthenings.
Another area in which additional study is needed is the type of bone formed at the site of the lengthening. Harris et al. recently reported on a patient with fibrous dysplasia in whom the bone in the lengthened segment had the same histological findings. It would have been useful to know the type of the regenerated tissue in our study, but this would have necessitated a biopsy, which we believed was an inappropriate and unacceptable risk for the patient. Harris et al. also described malignant transformation of the regenerated tissue to an osteosarcoma. As far as we know, this is the only reported case of such transformation, but it raises the issue of a possible causal relationship between lengthening and malignant degeneration of a benign tumor.
Unlike Bell et al., we found no relationship between the number and severity of complications and the percentage of lengthening or the duration of treatment in our series. However, those authors reported a greater mean increase in bone length (40 per cent in twenty-four tibiae) than we did. These findings suggest that there may be a threshold percentage of bone lengthening above which the number of complications increases. The complications were greater in number and severity when the patients had had simultaneous correction of an angular or rotational deformity than when only lengthening had been performed.
Our results suggest that the Ilizarov technique is effective in restoring normal limb length in patients who have an underlying bone disorder; however, the high rate of complications after lengthening of abnormal bone compared with that after lengthening of normal bone raises questions regarding the safety of this method and warrants meticulous selection of patients. Our present recommendation is that children who have a limb-length discrepancy should be evaluated with three consecutive scanograms to determine the estimated discrepancy at skeletal maturity. If this discrepancy is expected to be more than 15 per cent of the original limb length, limb-lengthening options should be discussed with the family. The timing of the procedure should be arranged to best suit the patient and the family.