Ultrasound, a form of mechanical energy that produces micromechanical
strain when transmitted through the body, is widely used in medicine as a
therapeutic and diagnostic tool. In vitro studies have demonstrated
that low-intensity pulsed ultrasound alters metabolism at the cellular level.
Micromechanical strain caused by ultrasound transmission through the body can
promote bone formation in a manner comparable with the bone responses to
mechanical stress postulated by Wolff's
law1,2.
Studies of animals have demonstrated that low-intensity pulsed ultrasound
increases the rate of endochondral bone formation and the structural strength
of fracture
sites3,4.
Clinical investigations have shown that low-intensity pulsed ultrasound
improves healing of pseudarthroses, delayed unions, and
nonunions5,6.
Randomized, double-blind, controlled clinical trials have shown that the
technique accelerates the repair process in fractures of the
tibia7 and
radius8. These
studies have established the clinical usefulness of low-intensity pulsed
ultrasound for acceleration of
fracture-healing5.
Callotasis and hemicallotasis are useful techniques to correct limb
deformity or to lengthen the
limb9, but the
distraction callus heals very slowly, requiring patients to wear an external
fixator for a long period. The effects of low-intensity pulsed ultrasound on
healing of the distraction callus have been examined in animal
models10-14
but have not been assessed in a clinical setting, to our knowledge.
Opening-wedge high tibial osteotomy by hemicallotasis, a surgical
intervention for patients with osteoarthritis in the medial compartment of the
knee, has the advantage of accurate
correction15.
However, the technique requires external fixation until the distraction callus
is mature, and a high rate of pin-site infection has been
reported16.
In this randomized study, we quantitatively examined the effects of
low-intensity pulsed ultrasound on the maturation of the distraction callus
during the consolidation phase in patients treated with an opening-wedge high
tibial osteotomy by hemicallotasis.
Patient Selection
From 1999 to 2002, bilateral one-stage opening-wedge high tibial osteotomy
by hemicallotasis was performed in thirty-three patients who had bilateral
osteoarthritis in the medial compartment of the knee with bilateral knee pain.
Selection criteria for the present study included (1) bilaterally symmetric
grades of osteoarthritis according to the classification of Kellgren and
Lawrence17, and (2)
a difference of =3° in varus deformity between the left and right sides
as indicated by the femorotibial angles measured on full-length radiographs of
the lower limbs with the patient standing. Twenty-one patients (seventeen
women and four men) who met these criteria consented to participate in the
study. The ages of the patients ranged from fifty-three to seventy-eight years
(average, sixty-eight years). The severity of the medial osteoarthritis was
grade 2 in four knees, grade 3 in thirty knees, and grade 4 in eight knees.
The study was approved by our hospital review board.
Operative Technique
Opening-wedge high tibial osteotomy was performed with use of an
articulated dynamic axial fixator with a proximal T-clamp (hemicallotasis
device; Orthofix 20-010 and CP-0029; Orthofix, Bussolengo, Italy). During the
surgery, four tapered half-pins (Orthofix; 6 mm/5 mm) were inserted; two with
cancellous threads were placed at the proximal sites and two with cortical
threads, at the distal sites. For the osteotomy, an anteromedial approach was
used, with a transverse incision of the skin and a longitudinal incision of
the periosteum followed by elevation. The position and direction of the
osteotomy were monitored in both the coronal and the sagittal plane with use
of radiography during the operation. Osteotomy of the medial three-quarters of
the proximal part of the tibia was carried out just proximal to the tibial
tubercle. We used an osteotome to cut the anterior and middle portions of the
tibia, and we used an oscillating saw to cut the posterior cortex. The status
of the cortex lateral to the osteotomy site was assessed by testing the
rigidity of the osteotomy site with manual valgus stress. The Orthofix fixator
was applied, and distraction was performed for 5 mm to ensure that the intact
portion lateral to the osteotomy site acted as a hinge. Then the fixator was
compressed so that there was good osseous apposition. The periosteum was
closed. A fibular osteotomy was not performed. As with callus distraction for
limb-lengthening, the osteotomy was followed by periods of latency,
distraction, and consolidation until removal of the external fixator. Patients
were allowed full weight-bearing throughout these periods.
Distraction and Consolidation
After a latency period of fourteen days, distraction was begun. Radiographs
of the osteotomy sites in both limbs were made once a week throughout the
distraction and consolidation phases. Additional radiographs were made
according to the symptoms and findings of previous radiographs. The
compression/distraction unit was elongated at a rate of 1 mm daily, with two
0.5-mm distractions each day. Distraction was stopped and the consolidation
phase was begun when the hip-ankle mechanical axis of the limb passed the
medial one-ninth to one-third of the lateral compartment of the tibial plateau
as seen on a full-length radiograph of the limb with the patient standing. The
locking bolt of the telescopic body of the fixator was then tightened.
The consolidation period was ended when the callus was considered strong
enough for safe removal of the fixator without the risk of fracture (i.e.,
when there was a smooth uninterrupted cortical margin medial to the
regenerated bone or an uninterrupted trabecular pattern occupying the lateral
two-thirds of the tibia on an anteroposterior radiograph). The assessors of
callus consolidation were not blinded to the use of low-intensity pulsed
ultrasound. A few days after release of the locking screw (at a mean [and
standard deviation] of 1.5 ± 0.6 days; range, one to three days), the
fixator was removed and the pins were left in place for a trial period of
several more days. The pins were removed after radiographs confirmed that
there was no evidence of collapse of the callus.
Measurement of Bone Mineral Density
Although we performed the osteotomy perpendicular to the long axis of the
lower limb in the sagittal plane, the postoperative anteroposterior radiograph
of the proximal part of the tibia revealed that the osteotomy was not
perfectly parallel to the x-ray beam in some patients. Before measuring bone
mineral density, all patients were placed in the supine position with the
knees in neutral rotation under the x-ray tube. The knee joint was slightly
flexed (Fig. 1-A) or the entire
limb was raised (Fig. 1-B), so
that the beam would pass parallel to the osteotomy. As a result, we obtained
an anteroposterior view of the callus without an overlap of the anterior or
posterior cortical bone of the tibia. We carefully recorded the position of
each limb of each patient by measuring the distance and angle of the fixators
relative to the bed on which the patient lay. To measure the bone mineral
density of the distraction callus, the patient was then placed on a
dual-energy x-ray absorptiometry unit (QDR-2000; Hologic, Boston,
Massachusetts) with the limb in the same position as determined under the
x-ray tube. Bone mineral density was expressed in grams per square centimeter.
On the view obtained from the scan, we confirmed that the callus was visible
without overlap of anterior or posterior cortical bone of the tibia,
indicating that the beam was parallel to the osteotomy. Two regions of
interest in each tibia were selected for measurement of regional bone mineral
density with use of QDR for Windows software (version 11.2; Hologic).
Following the previously described method with
modification18, we
selected one region of interest within the distraction gap
(Fig. 1-C). The other was in
the segment just distal to the distraction gap
(Fig. 1-D). The region of
interest in the distal segment was 8 mm in height and spanned the medial half
of the tibia. The reproducibility error of the results was evaluated with
three consecutive measurements and was found to be <3%. To test day-to-day
reproducibility, a bone phantom (Hologic) was attached with a metal pin 5 cm
above the bed, and its bone mineral density was measured daily for one month.
The coefficient of variation was 0.8%.
The bone mineral density of the callus was measured twice for each tibia:
at the start of the consolidation period and four weeks after the start of the
consolidation period. The first bone mineral density measurement was
subtracted from the second to calculate the increase in bone mineral density
over the four-week consolidation period.
Application of Ultrasound
For each patient, the limb to be treated with low-intensity pulsed
ultrasound was randomly selected with use of a random-number generator on a
computer. Ultrasound energy was provided by a Sonic Accelerated Fracture
Healing System (SAFHS; Exogen, Piscataway, New Jersey). The treatment head
module delivered an ultrasound signal composed of a burst width of 200 µsec
containing 1.5-MHz sine waves, with a repetition rate of 1 kHz and a spatial
average-temporal average intensity of 30 mW/cm2. Within two days
after the first bone mineral density measurement, we began daily twenty-minute
ultrasound treatments on the one side, and we continued the treatment until
the fixator was removed. The treatment head module was positioned on the
anteromedial aspect of the proximal part of the leg at the level of the
osteotomy and was fixed with a strap. All twenty-one patients were
hospitalized until removal of the pins and were assisted by the hospital staff
in carrying out the low-intensity pulsed ultrasound treatment.
Statistical Analysis
The t test was used to compare data between treated and untreated limbs. A
p value of <0.05 was considered to indicate significance.
Baseline Characteristics
None of the patients had previously undergone knee surgery such as
meniscectomy or cruciate ligament repair. The mean weight of the patients was
61 ± 11 kg, and the mean body mass index was 26 ± 3. Three
patients smoked during the treatment period.
Preoperatively, the femorotibial angle averaged 185° ± 5° in
the ultrasound-treated limbs and 184° ± 5° in the controls, the
range of knee flexion measured with manual goniometry averaged 121°
± 27° in the ultrasound-treated limbs and 125° ± 10°
in the controls, and the range of knee extension averaged -4° ±
6° in the ultrasound-treated limbs and -3° ± 5° in the
controls. Thus, the preoperative varus deformity and ranges of motion did not
differ substantially between the ultrasound-treated and control limbs.
Preoperatively, four patients had equal pain in the two knees, seven had more
knee pain on the side that was later treated with low-intensity pulsed
ultrasound, and ten patients had less pain on that side. The mean period of
distraction did not differ between the ultrasound-treated tibiae (37 ±
7 days) and the control tibiae (37 ± 7 days). The distraction period
was the same for both sides in eighteen patients. It was two days shorter on
the ultrasound-treated side in one patient, four days shorter on the
ultrasound-treated side in another patient, and three days shorter for the
control limb in still another patient.
During the distraction period, patients experienced pain in the lower limbs
and were occasionally unable to stand. After the start of the consolidation
period, the pain decreased, and all patients walked using a pair of crutches.
Pin-track infections developed in six limbs in the ultrasound-treated group
and five limbs in the control group. These infections responded to local
pin-site care and antibiotic treatment. We did not observe signs of infection
at any osteotomy site or in any joint in the limbs.
As a result of randomization, the right limb was chosen for low-intensity
pulsed ultrasound and the left limb served as the control in eleven patients.
In ten patients, the left limb was chosen for low-intensity pulsed ultrasound
and the right limb served as the control.
Bone Mineral Density
Before the start of the treatment with the low-intensity pulsed ultrasound,
the bone mineral density of the distraction callus averaged 0.30 ± 0.11
g/cm2 in the tibiae chosen to be treated with ultrasound and 0.32
± 0.09 g/cm2 in the controls (p = 0.31, unpaired t test)
(Fig. 2). The mean increase in
bone mineral density during the four-week treatment period was significantly
greater in the ultrasound-treated tibiae (0.20 ± 0.12 g/cm2)
than it was in the controls (0.13 ± 0.10 g/cm2) (p = 0.02,
unpaired t test). In eighteen patients, the bone mineral density increased
more in the ultrasound-treated limb than in the control limb; in three
patients, the bone mineral density increased more in the control limb (Figs.
3-A,
3-B,
3-C,
3-D,
4). Of the three patients who
had a greater increase in the control limb, one was a smoker, two had had
almost equal pain in the two knees, and one had had more knee pain in the
ultrasound-treated limb than in the control limb.
The mean increase in bone mineral density in the segment just distal to the
distraction gap during the four-week treatment period was 0.02 ± 0.09
g/cm2 in the ultrasound-treated tibiae and —0.03 ±
0.09 g/cm2 in the control tibiae. The difference between the two
groups was not significant (p = 0.07).
Consolidation Period
The mean consolidation period was 7.1 ± 2.6 weeks (range, four to
twelve weeks) in the ultrasound-treated group and 7.9 ± 2.4 weeks
(range, five to thirteen weeks) in the control group. The consolidation period
for the ultrasound-treated limb was shorter than that for the control limb in
thirteen patients: it was three weeks shorter for four patients, two weeks
shorter for four patients, and one week shorter for five patients. The
consolidation period was the same on both sides in six patients, and, in two
patients, consolidation occurred one week earlier in the control limb than in
the ultrasound-treated limb. The fixators remained in place for a mean of 14.6
± 3.0 weeks in the ultrasound-treated tibiae and for 15.5 ± 2.7
weeks in the controls.