Selection Criteria
Our hospital hosts the regional malunion/nonunion referral clinic.
Recruitment of subjects for the present study was undertaken following a
review of the clinical notes on the forty-one patients who had undergone
reconstruction of a distal tibial defect with use of distraction osteogenesis
techniques at our institution within the previous five years. To be included
in the study, a subject had to have a fully functional and uninjured
contralateral lower limb that could act as a control, had to have been
weight-bearing during the treatment with distraction osteogenesis, had to have
returned to full employment within the first year following the removal of the
device, had to have had the device removed at least eighteen months prior to
the time of the study, had to be satisfied with the final outcome, and had to
have a range of motion at the ankle and knee joints that did not limit any
activities. The strictness of the inclusion criteria was necessary in order to
exclude patients with extensive muscular atrophy and neuropathic pathology so
that our cases would reflect an optimal result. An effort was made to select
patients with different durations of treatment, lengths of regenerated bone,
and durations of follow-up as it was thought that this would identify
abnormalities that characterize all patients during the late rehabilitation
period following bone transport for reconstruction of a tibial defect and not
reflect findings that are specific to only one group of such patients during a
specific period in time.
The fifteen most suitable patients were selected on the basis of the
clinical notes according to the above criteria, and they were invited for an
interview and clinical examination. For the purposes of this study, subjective
satisfaction was measured with a 10-point visual analogue scale, and patients
were included in the study if they had a score of >8 points. Following the
examination, only eight of the fifteen patients were considered to be suitable
for the project, as the remaining patients had undocumented residual
pathological findings involving the injured or contralateral limb or were
dissatisfied with the final outcome. The study received approval from the
local ethical committee authority.
Patient Characteristics
Seven of the volunteers had been treated for a bone defect due to
management of osteomyelitis following internal fixation of a distal tibial
fracture. Three of these patients had sustained an open fracture, and two of
them (Cases 5 and 7) had required split-thickness skin-grafting. The eighth
subject (Case 4) had had excision of a fibrous dysplasia lesion of the distal
part of the tibia. Details on the patients are presented in
Table I.
The treatment protocol was similar for all of the patients with
osteomyelitis. Initially, an external fixator was applied to preserve the
length and orientation of the tibia, following which the diseased segment was
excised. Subsequently, once there was clinical and hematological evidence that
the infective process had resolved (as indicated by the erythrocyte
sedimentation rate and serum C-reactive protein level), the patient underwent
a proximal corticotomy and appropriate alterations of the frame to facilitate
bone transport. In the patient with fibrous dysplasia (Case 4), the excision
of the pathological focus and the corticotomy were performed during the same
session. None of the patients had more than 1.5 cm of lower-limb length
discrepancy preoperatively, and in all cases the final length of the tibia was
equal to that of the contralateral, uninjured tibia. No patient had any pins
inserted into the foot; therefore, free mobility of the ankle joint was
allowed throughout the treatment period. Four patients (Cases 2, 4, 7, and 8)
had slow mineralization of the regenerated bone, and in two of them (Cases 4
and 8) a second frame had to be applied because of nonunion of the docking
site. Full weight-bearing was encouraged throughout the treatment period. All
patients were asked to attend a weekly dedicated physiotherapy outpatient
clinic. In addition, additional physiotherapy sessions were arranged depending
on individual needs.
Follow-up had been completed for all patients, and at least nineteen months
(mean [and standard deviation], 39.5 ± 9.1 months) had elapsed since
the removal of the external fixator. During the examination for this study,
all patients were pain-free and were able to walk and climb stairs without
difficulty. Two of the patients (Case 5 and 8) were observed to have a minor
limp, whereas the others walked without any noticeable limitations. Only two
patients (Cases 1 and 4) stated that they could run and participate in sports
activities. Two other patients (Cases 6 and 7) reported that they were not
involved in physically demanding activities, which they indicated was partly
because of their age.
On clinical examination, four of the patients had the same amount of
extension of the two knees, while the remaining four had a fixed flexion
deformity of up to 8°, as measured clinically with a goniometer, of the
involved knee. In all cases, the ankle could be placed passively in the
plantigrade position while the knee was flexed (mean ankle dorsiflexion,
10.7° ± 4.3°). With the knee fully extended, however, the mean
ankle dorsiflexion was limited to 4.7° ± 5.6°, indicating that,
on the average, the patients had 6° of tightness of the gastrocnemius. All
of these measurements were performed by a single investigator (see
Appendix).
On the basis of previously published criteria for judging the final outcome
of similar
conditions17, six
of the patients were considered to have an excellent outcome and the remaining
two (Cases 5 and 8) had a good outcome. All but two patients (Cases 2 and 4)
were smokers.
Gait Analysis
Seven of the eight patients underwent gait analysis. For comparison
purposes, gait data were also collected from a control group consisting of
seven healthy adults, with a mean age (and standard deviation) of 26.1
± 9.4 years, a mean height of 179.0 ± 9.1 cm, and a mean body
mass of 84.0 ± 12.2 kg.
A five-camera three-dimensional motion analysis system (Vicon 250; Oxford
Metrics, Oxford, United Kingdom) was used to record (at 50 Hz) the gait of all
subjects (patients and controls) as they walked across two force-platforms
(AMTI OR6-7; Advanced Mechanical Technologies, Boston, Massachusetts) mounted
in the middle of a 10-m walkway. Subjects walked barefooted at their
self-selected, normal walking speed. Only trials in which each foot was placed
within the bounds of a single platform without understriding or overstriding
were saved for further analysis. Each patient completed ten trials, whereas,
because asymmetry between sides was not investigated in the control subjects,
the controls were required to complete only three trials each.
With use of the Plug-In Gait software (Vicon Motion Systems, Oxford, United
Kingdom), three-dimensional marker trajectory data were filtered and then were
processed to produce a three-dimensional link-segment model of the lower
limbs. Subsequent data analysis was performed with use of Polygon software
(Vicon Motion Systems). The data from the ten repeated trials were combined to
produce an average file for each patient. Stance time, step length, and the
angular displacements, moments, powers, and work done at the ankle were then
determined for both the affected and the unaffected side.
The data collected from all control subjects were combined to produce a
mean file for the control group. Because gait is highly individualized, it is
difficult to make reference to an ideal or optimum gait pattern. One could
argue, therefore, that a control-group average has limited meaning, and this
certainly would be the case if the control group was being used to make
statistical comparisons. The gait data collected from the control subjects
were determined in an attempt to highlight, qualitatively (i.e., graphically
[Fig. 2]), the joint kinematic
and kinetic alterations that may have occurred as a result of the bone
transport. Without such a comparison, it would be difficult to determine
whether the kinematic and kinetic data determined for the unaffected limbs
were typical of those seen in normal gait.
Dynamometry
In order to measure the moment-generating potential of the muscles that
span the ankle, all eight volunteers were examined with use of a dynamometer
(Cybex Norm; Cybex, division of Lumex, Ronkonkoma, New York) set in isometric
mode. Each subject was placed in the prone position when the plantar flexor
muscles were tested and in the supine position when the dorsiflexor muscles
were tested. The knee was fully extended and the foot was secured in its
neutral position (plantigrade) on the dynamometer foot-plate with inelastic
strapping. When a patient had minor fixed flexion deformities at the knee
and/or ankle joints, the unaffected limb was positioned appropriately to
ensure examination of a similar joint configuration in both limbs. The
affected and unaffected limbs were examined in random order.
After the subjects had been sufficiently familiarized with the test
protocol, they were requested to produce isometric maximum voluntary
contractions in the plantar flexion and dorsiflexion directions. In each test,
they were asked to increase their effort gradually within a period of two to
three seconds and to maintain their maximal effort for an additional second.
Five tests were performed for each muscle group with a rest period of two
minutes between tests. The test that produced the highest joint moment was
included in the analysis. None of the examinations were limited to any degree
by pain or discomfort as judged by the patients.
Electromyographic Measurements
Ten-millimeter-diameter bipolar Ag-AgCl surface electrodes (Medicotest,
Rugmarken, Denmark) with a center-to-center distance of 2 cm were used to
measure the electromyographic activity of the tibialis anterior and triceps
surae muscles during the maximum voluntary contractions. Before electrode
placement, the skin was shaved and was cleaned with alcohol to reduce
impedance. For the measurements of the tibialis anterior muscle, the
electrodes were placed along the muscle belly about 10 cm distal to the knee
joint line. For the measurements of the triceps surae muscle, a weighted
recording of the maximum voluntary contractions was collected by placing one
electrode on the medial head of the gastrocnemius muscle and the other on the
soleus muscle. In all cases, the neutral electrode was placed over the lateral
epicondyle of the femur of the tested limb. The electromyographic signals were
recorded by amplifiers (Biopac Systems, Santa Barbara, California) with a gain
of 1000 and a frequency band of 10 to 500 Hz. After digital conversion of the
signal, full-wave rectification and integration over 50-msec intervals was
carried out. Integrated electromyographic signals were averaged over a
one-second period corresponding to constant maximal joint moment
production.
Statistical Analysis
Results are presented as the mean and standard deviation. The different
parameters from the gait, dynamometric, and electromyographic analyses were
compared between the affected and unaffected limbs with use of the paired
Wilcoxon test.
Gait Analysis
Figure 2 shows the
mean angular displacement, moment, and power of the affected and unaffected
ankles of a typical patient (Case 4) as well as the mean values and standard
deviations for the control group (gray shaded area). It is evident that there
were considerable kinematic and kinetic differences between the affected and
unaffected limbs of the patient as well as between the affected limb of the
patient and the control group mean. It is also evident that the data for the
unaffected limb were similar to those for the control group. Although it is
possible that the unaffected limb may have compensated in some way for the
reduced function on the affected side, the data tend to suggest that kinematic
and kinetic alterations at the ankle joint were evident only in the affected
limb. These alterations are described quantitatively below.
All subjects had similar step lengths on the affected and contralateral
sides (p = 0.91), despite the stance time being consistently shorter for the
affected limb (0.43 ± 0.003 sec) compared with the unaffected limb
(0.47 ± 0.003 sec) (p = 0.01).
Plantar flexion immediately following heel-strike varied in the study
population, with a mean of -7.5° ± 2.7° in the unaffected limbs
and 10.1° ± 4.5° in the affected limbs (p = 0.12). However, the
mean dorsiflexion angle of the ankle during mid-stance was decreased in the
affected limb (mean, 12.1° ± 4.2° in the unaffected limbs and
8.2° ± 5.6° in the affected limbs; p = 0.06). The maximum
plantar flexion angle at toe-off was similar between the two limbs (mean,
-15.5° ± 6.7° for the unaffected limbs and -14.4° ±
5.5° for the affected limbs; p = 0.73). The angular displacement between
the first and the second rocker was also similar between the two limbs (mean,
-19.60° ± 3.80° for the unaffected limbs and -18.3°
± 2.1° for the affected limbs; p = 0.23), but the angular
displacement between the second rocker and the instant of maximum plantar
flexion was decreased in the affected limbs (mean, 27.7° ± 5.2°
for the unaffected limbs and 22.5° ± 3.3° for the affected
limbs; p = 0.04).
Peak dorsiflexion moment following heel-strike was similar between the two
limbs (mean, -0.06 ± 0.06 Nm/kg for the unaffected limbs and -0.05
± 0.06 Nm/kg for the affected limbs; p = 0.45). However, peak plantar
flexion moment during terminal stance was considerably decreased in the
affected limbs compared with the unaffected limbs (mean, 1.64 ± 0.25
Nm/kg for the unaffected limbs and 1.38 ± 0.19 Nm/kg for the affected
limbs; p = 0.04). Peak power in the affected limbs was decreased both in the
eccentric phase of stance (mean, -0.98 ± 0.63 W/kg in the unaffected
limbs and -0.66 ± 0.19 W/kg in the affected limbs; p = 0.03) and in the
concentric phase of stance (mean, 2.58 ± 0.63 W/kg in the unaffected
limbs and 1.73 ± 0.25 W/kg in the affected limbs; p = 0.01). The energy
absorbed during stance (negative work) was similar between the two limbs
(mean, -0.35 ± 0.10 J/kg in the unaffected limbs and -0.30 ±
0.09 J/kg in the affected limbs; p = 0.23), but the energy that was generated
during the second half of stance (positive work) was significantly decreased
in the ankles of the affected limbs (mean, 0.38 ± 0.09 J/kg in the
unaffected limbs and 0.26 ± 0.06 J/kg in the affected limbs; p =
0.02).
Dynamometric and Electromyographic Analysis
The maximum voluntary contraction moment generated in the plantar flexion
direction by the reconstructed limb was a mean of 27.4% ± 15.1% lower
than that generated by the unaffected limb (p < 0.012). Similarly, the
maximum voluntary contraction moment generated in the dorsiflexion direction
by the affected limb was a mean of 51.7% ± 17.8% lower than that
generated by the unaffected limb (p = 0.01). The data for each subject are
presented in Table II.
The electromyographic activity of the tibialis anterior and triceps surae
muscles of the affected limbs was also significantly reduced compared with
that of the unaffected limbs during maximum voluntary contraction in both
plantar flexion and dorsiflexion. A summary of the results is presented in
Table III.
We investigated the long-term functional outcomes of patients in
whom a segment of bone had been transported distally to reconstruct a distal
tibial defect. In all patients, the corticotomy was performed proximally
within the region where the tibialis anterior and soleus muscles arise. During
the transport process, a segment of the muscles remained attached to the
proximal part of the tibia while the remaining segments followed the
transported section of bone and migrated distally. To the best of our
knowledge, apart from clinical series, there have been no animal or human
research investigations of the response of muscle to shortening induced by
bone migration. One would expect, however, that some of the changes that have
previously been identified in the early period of distraction in patients
treated with lengthening would apply to the patients described here. It is
also possible that some intramuscular injury occurred during migration as the
muscle was sheared into two parts. Furthermore, there is little doubt that
some of the findings in the present study were the result of the traumatic
event that led to the initial fracture and possibly were also the result of
the iatrogenic injury that occurred during the surgical procedures.
Gait analysis was undertaken to determine to what extent each patient had
regained a "normal," symmetrical gait pattern. By means of a
qualitative comparison with data collected from a control group of healthy
subjects, the gait analysis also demonstrated the extent, if any, to which the
unaffected limb of the patients had compensated for the affected limb. There
were considerable kinematic and kinetic differences between the affected and
unaffected limbs as well as between the affected limbs and the control group.
Furthermore, the data for the unaffected limbs were similar to those for the
control group. The data suggested that kinematic and kinetic alterations at
the ankle joint were evident only in the affected limbs, as discussed
below.
Stance time on the affected limb was significantly shorter than that on the
unaffected limb (p = 0.01). While this was a common finding, it was more
pronounced in the two patients who walked with the ankle in equinus. This
finding could be the result of weakness of the gastrocnemius or of the
existence of the equinus contracture, as previously described in patients who
had undergone correction of a limb-length
discrepancy20. The
patients in our study had a decrease in the ankle angular displacement of the
affected limb during the first half of stance. During this period, the
tibialis anterior muscle initially lengthens, whereas it contracts to control
ankle plantar flexion as the foot is lowered to the ground following heel
contact; then the triceps surae muscle contracts eccentrically, stabilizing
the ankle as the center of gravity moves anteriorly over the planted foot.
During the second half of the stance period, from the point of maximum
dorsiflexion to push-off (i.e., maximum plantar flexion), the angular
displacement at the ankle of the affected limb was consistently less than that
of the ankle of the unaffected limb. During this period, the triceps surae
muscle and the flexors of the first ray contract to displace the body forward.
These findings indicate a reduced ankle moment and power in the involved limb,
particularly in the second half of stance up to the time of push-off. As none
of the patients had any functional restriction of plantar flexion mobility,
the reduced ankle function was either a result of a change in muscle function
due to its surgical manipulation or a result of the equinus, which altered the
way the foot was loaded during stance.
Even though these results provide some insight into the gait pattern of
patients following reconstruction of the tibia, they should be considered with
caution. Gait can be substantially affected not only by changes in the muscle
but also by alterations in the mobility of the joints. Indeed, all of the
patients in our study presented with a degree of contracture of at least one
of the joints adjacent to the tibia. These contractures inevitably led to
compensatory changes and thus may explain, at least in part, the differences
found between the affected and unaffected limbs.
However, the possibility that the lower-limb muscles of the operatively
treated limb had undergone substantial functional changes is supported by the
dynamometric and electromyographic data. The dynamometry tests were performed
isometrically at a given joint configuration and thus the minor fixed flexion
deformities noted in some of the patients should not have affected the
outcomes of these tests. The substantially decreased moment-generating
potential and associated electromyographic activity in the limbs that had had
the reconstruction compared with the unaffected limbs may explain the
reluctance of most of the patients to participate in more demanding physical
activities such as running or sports and may explain the differences found
during the gait analysis. Muscle size reduction may account for the reduction
in the maximum voluntary contraction moments in the affected limbs. This
explanation is supported by experimental evidence that immobilization induces
skeletal muscle atrophy, reduces maximum voluntary contraction moments, and
reduces electromyographic activity during maximum voluntary
contractions20-23.
Loss of muscle could also explain why the decrease in the maximum voluntary
contractions of the affected limb compared with those of the unaffected limb
was greater during dorsiflexion (50.7%) than during plantar flexion (27.4%),
as the tibialis anterior can be more vulnerable to injury in a tibial fracture
as well as during its surgical exposure than are the plantar flexors. The
decrease in the electromyographic readings may have resulted from a loss of
motor units or from alterations of their characteristics, such as those
occurring with changes in muscle fiber type or a decrease in the neural
drive24. An
emerging theory that could explain the findings in the present study is that
there is loss of motor units, possibly due to denervation injury and cell
death during the transport process, in combination with alteration of the
electrical activity of the remaining motor units, possibly due to muscle fiber
type transformation.
All of the cases presented in this report demonstrated striking
similarities. One might have expected that the noted effects would have been
more pronounced as the migration distance of the transported segment
increased. The limited number of participants did not permit sensible
statistical analysis in order to demonstrate this. Similarly, we were unable
to identify any trends relating the outcome to the time between the
development of the pathological condition and the treatment, the duration of
treatment with the external fixator, or the time between removal of the
fixator and the investigation.
In conclusion, the findings of the present study offer an objective measure
of the long-term functional outcomes of patients in whom a segment of bone had
been transported distally to reconstruct a distal tibial defect. Although we
were unable to provide a precise explanation for the adaptive processes, what
was certain was that the affected limb had undergone considerable changes that
affected both routine and maximal effort capabilities. To help improve the
clinical outcome in these difficult cases, additional research is required in
order to understand the exact mechanisms involved.