The concept of multiple osteotomies and internal fixation with an
intramedullary rod for deformity correction and prevention of long-bone
fractures in patients with osteogenesis imperfecta was first described by
Sofield and Millar in
19591 and has been
widely accepted. In 1963, Bailey and
Dubow2 introduced an
elongating rod system, whereby a hollow sleeve and an internal obturator were
engaged and anchored by T-pieces at the proximal and distal epiphyses of the
long bones. This design allowed telescoping elongation as the long bone grew
at the physes, and effectively decreased the number and frequency of
operations required in growing
children3-8.
However, its use was technically demanding, and high rates of operative and
postoperative complications were
reported6,9-15.
The Sheffield telescopic rod was devised to eliminate problems associated with
the T-piece10 but
was nevertheless associated with other complications, such as intra-articular,
metaphyseal, or extracortical rod migration. Injuries caused by the
implantation of elongating rods with a T-piece are substantial but are rarely
discussed. Insertion of the distal obturator is more invasive than is
insertion of the proximal sleeve in both the femur and the tibia, as it
requires an arthrotomy of the distal joint, which may damage the articular
cartilage5,8.
Installation of the obturator with a T-piece into the tibia is even more
traumatic, requiring complete dislocation of the ankle joint or a medial
malleolar
osteotomy5,8,
possibly damaging the weight-bearing area of the distal tibial articular
cartilage. These possible iatrogenic problems along with higher complication
rates13,16
have made use of the telescopic rod in the tibia controversial, and as a
result some authors prefer rods that do not
elongate5,8.
We developed a new anchorage system to be used in the distal epiphysis to
allow telescoping of an elongating rod. This system does not require
arthrotomy for insertion of the obturator, and thus damage to ligaments and
articular cartilage can be avoided. The system was originally devised for the
tibia, but with accumulated experience its use was extended to the femur. The
purposes of this study were to introduce this new design and to report the
intermediate follow-up results of its use in patients with osteogenesis
imperfecta.
Implant Design
The interlocking telescopic rod is composed of a hollow sleeve and an
internal obturator. The sleeves are of the same shape and dimension as those
used in the telescopic intramedullary rod system (Sheffield Rod; Downs
Surgical, Sheffield,
England)10,17.
The obturators (Dyna-Locking Telescopic Rod; U&I, Gyeonggi, Korea) have
the same diameters as those of the Sheffield rod system. The interlocking
telescopic rod does not have a T-piece at the distal end but instead has a
small hole that allows the placement of an interlocking pin (a Kirschner wire,
with a threaded tip, of 1.4 mm in diameter [for 3, 4, and 5-mm-diameter
sleeves] or 1.8 mm in diameter [for 6 and 8-mm-diameter sleeves])
(Fig. 1). The obturators pass
all of the way through the corresponding sleeves except for the 3-mm-diameter
sleeve, for which the obturator has a widened distal tip to receive a
1.4-mm-diameter interlocking pin.
Surgical Technique
Preoperative radiographs are thoroughly evaluated to assess the deformity
and to estimate the lengths and diameters of the rods to be used. However, for
severely angulated limb segments, we determined the rod length
intraoperatively, after performing multiple osteotomies.
The surgical procedure (Figs. 2-A and
2-B) consists of multilevel osteotomies performed to realign the
limb segment along the rod. Percutaneous
osteotomies8 are
preferred, but open osteotomies are performed if intramedullary reaming is
needed for a narrow or obliterated medullary cavity or resection of a
substantial wedge of bone is needed to correct an acute angulation. A
Kirschner wire is then inserted through the medullary canal of the osteotomy
fragments. The sleeve is cut to an appropriate length and is inserted over the
Kirschner wire in an antegrade direction. It is important to ensure that the
distal tip of the sleeve points to the center of the distal epiphysis on both
anteroposterior and lateral projections of radiographs. The Kirschner wire is
then replaced with the obturator, which is advanced antegrade inside the
sleeve into the distal epiphysis. The rotational orientation of the hole at
its distal end can be controlled and adjusted with use of an
obturator-impactor (Fig. 1).
With use of a free-hand technique with fluoroscopic imaging, the obturator is
then anchored in the distal epiphysis with a Kirschner wire. As the cortex of
the distal epiphysis is thin in patients with osteogenesis imperfecta, the
Kirschner wire can be inserted manually and then advanced by gentle tapping.
The wire is cut to an appropriate length and is pushed deep, preferably within
the osseous epiphysis. Finally, in the femur, the sleeve position is adjusted
so that the T-piece abuts the greater trochanter cartilage within the gluteal
muscles. In the tibia, the T-piece of the sleeve is buried within the osseous
epiphysis. In this series, it was rotated 90° in ten cases as recommended
by Marafioti and
Westin3, but it was
placed only subchondrally to facilitate future rod replacement in thirteen
cases.
In the case of the 3-mm sleeve, through which the obturator cannot pass
because of its widened distal tip, the sleeve and obturator are assembled and
then inserted as one piece through the osteotomy fragments. When the sleeve
reaches its destination, the obturator is further advanced into the distal
epiphysis and is transfixed with use of the technique described above.
All limbs are immobilized with a long leg splint for four to six weeks
postoperatively.
Patients and Evaluation
This study was approved by our institutional review board, and the patients
and/or their parents gave informed consent to participate. The first fifteen
consecutive patients (seven girls and eight boys) with osteogenesis imperfecta
who had undergone insertion of an interlocking telescopic rod into the femur
or tibia and had been followed for more than two years after the index
operation were included. Three patients had osteogenesis imperfecta type I
according to the classification system of Sillence et
al.18, three had
type III, and nine had type IV.
Twenty-three tibiae and nine femora underwent surgery. The patient's age at
the time of the surgery and the follow-up period were recorded separately for
each operatively treated limb segment. The mean age at the time of the surgery
was 7.3 years (range, 1.9 to 11.9 years), and the mean duration of follow-up
was 3.3 years (range, 2.0 to 6.9 years). Follow-up was considered to be
complete when any part of the rod system was removed or when skeletal maturity
was reached. All patients received a bisphosphonate, either oral alendronate
or cyclic intravenous pamidronate, before and after the index operations. The
patients were examined clinically every three to six months, and radiographs
were made every six to twelve months. The fracture history and complications
were recorded.
Survival analysis with the Kaplan-Meier method was performed with use of
SPSS for Windows software (version 12.0; SPSS, Chicago, Illinois). In the
calculation of surgery-free survival, any additional surgical intervention
such as removal or replacement of the interlocking telescopic rod, advancement
of the proximally migrated sleeve, or adjustment of the interlocking pin was
defined as the end point. In the calculation of rod survival, the end point
was defined as removal or replacement of the rod due to a complication. As
long as a rod remained in the operatively treated limb segment, regardless of
whether it was elongating, it was considered to have not reached the end
point.
Application of the intramedullary rod and the interlocking pin fixation
were performed successfully in all limb segments, and telescoping took place
uneventfully in all of the segments for at least eighteen months (Figs.
3 and
4).
Five operatively treated limb segments sustained a fracture
postoperatively: one femur showed a stress fracture, and the other femur and
the three tibiae sustained a linear fracture following minor trauma. All
healed uneventfully after a brief period of immobilization with a splint.
In one tibia, in which the interlocking pin had been placed eccentrically
at the anterior part of the distal epiphysis, the rod cut through the anterior
cortex of the distal part of the tibia. Gradual anterior angulation at the
distal part of the tibia resulted in the need for a closing anterior wedge
osteotomy and insertion of another interlocking telescopic rod four years
postoperatively. Proximal migration of the sleeve was observed in two limb
segments. One tibial sleeve, which had been implanted subchondrally, migrated
proximally into the knee joint. It was repositioned through a small arthrotomy
and then remained subcortical for the next two years. One femoral sleeve
migrated proximally and was relocated and tied to the proximal part of the
femur with a wire.
At the time of the latest follow-up, seven patients could walk without
limitation, six could walk a short distance without a walking aid but required
an aid for long distances, one required a walking aid at all times, and one
could stand and walk only during rehabilitation sessions.
Backing out of the interlocking pin was observed, at five to thirty-three
months postoperatively, in the first four tibiae and the first femur treated
in this series. In these five cases, a smooth Kirschner wire had been used as
the interlocking pin. In three of the tibiae, the backed-out interlocking pin
was removed, allowing the rod to function as a non-elongating rod. One of
these patients reached skeletal maturity without the development of any
angulation or fracture of the tibia, whereas the other two patients underwent
removal of the rod to treat progressive tibial angulation at the tip of the
rod or osteomyelitis. In the remaining two cases, the interlocking pin was
replaced with a pin with a threaded tip, which was inserted deep into the
epiphysis, as soon as backing out was detected. The new pin did not back out
in either of the two cases, which were followed for three and four years. None
of the twenty-seven limb segments in which a Kirschner wire with a threaded
tip had been used initially had backing out at the time of follow-up, at an
average of 3.1 years.
Proximal migration of the obturator with the interlocking pin cutting
through the physis was observed in four tibiae in two patients. Both had
Sillence type-IVA osteogenesis imperfecta, and they underwent the index
operation at the ages of three and 4.5 years. The proximal migration was
detected at 2.5 and three years postoperatively, as the rods failed to
telescope even though they appeared straight. In one patient, both
interlocking pins were repositioned within the distal epiphysis, after which
the sleeves started to migrate distally. In the second patient, no additional
surgery was done and the intramedullary rods functioned as non-elongating rods
(Fig. 5).
Twenty-three (72%) of the thirty-two rods telescoped for an average of 2.8
years without any complications. The cumulative surgery-free survival rate of
the rods was 90.6% (95% confidence interval, 80.5% to 100.7%) at two years,
71.6% (95% confidence interval, 53.1% to 90.0%) at three years, and 58.7% (95%
confidence interval, 36.4% to 81.0%) at four years
(Fig. 6). The rod had to be
removed or replaced in only three cases. The cumulative survival rate of the
rod itself was 88.7% (95% confidence interval, 73.7% to 103.6%) at four years
(Fig. 6).
Multiple osteotomies with intramedullary fixation are the mainstay of
treatment of the long bones of growing children with osteogenesis imperfecta.
We believe that the use of an elongating rod is beneficial in patients with
substantial remaining growth. However, the Bailey-Dubow elongating nail has
been associated with a high complication rate, particularly related to the
T-piece6,9-15.
The Sheffield telescopic rod system, a modification of the Bailey-Dubow nail,
eliminated T-piece-related complications by permanently fixing the T-piece to
the
sleeve10,17,
but it shares the problems of traumatic insertion and rod migration with its
predecessor19,20.
Placement of an obturator with a T-piece at the distal fragment is an
invasive procedure, especially in the tibia. It requires almost complete
dislocation of the ankle joint with use of either extensive division of the
deltoid ligament or medial malleolar
osteotomy7,11,17,19,20.
Also, the T-piece has to be inserted through the weight-bearing surface of the
distal tibial articular cartilage, which leaves a large scar in a relatively
narrow area. Some
authors3,9,20
have reported no functional change of the ankle joint after ankle arthrotomy,
but damaged articular cartilage and subsequent fibrocartilaginous healing are
highly likely to have a detrimental effect on the ankle joint in the long run,
especially in patients who can walk. Therefore, installation of the telescopic
rod with a T-piece remains controversial for tibial
bones8.
Janus et al.15
described a modification of the Sheffield rod system for the tibia in which a
prebent obturator was anchored at the medial malleolus, eliminating the need
for an arthrotomy of the ankle joint. However, ankle valgus deformity due to
an inadequately prebent obturator was observed in some cases. With use of the
interlocking telescopic rod reported on here, not only can ankle joint
arthrotomy and damage to the distal tibial articular cartilage be avoided, but
accurate anatomic alignment of the tibia can be restored.
Another problem associated with distal tibial fixation is that often the
distal tibial epiphysis in patients with osteogenesis imperfecta is
hypoplastic compared with the proximal epiphysis. As the distal fixation pin
of the interlocking telescopic rod is thinner than the T-piece, the
interlocking telescopic rod can be used in younger children. In the current
series, the youngest patient treated with a tibial interlocking telescopic rod
was 3.2 years old.
It is not uncommon for patients with osteogenesis imperfecta to require
revision surgery as they age. Sometimes an obturator with a T-piece is
impossible to remove without causing serious damage to the joint or
surrounding tissue7.
In the interlocking telescopic rod system, the obturator can be converted into
a smooth rod simply by removing the interlocking pin. Then the obturator can
be easily removed through the diaphyseal medullary cavity or through the
articular cartilage with minimal injury.
The concept of interlocking the telescopic rod at the distal part of the
tibia is not new.
Bailey4 reported a
case in which the sleeve was placed distally and then transfixed with use of a
cross-wire passed through the epiphysis. However, this configuration failed as
the interlocking pin backed out. The smooth interlocking pins that we used
initially, in the first four tibiae and the first femur that we treated,
backed out, but this problem was overcome by using a Kirschner wire with a
threaded tip and inserting it deep into the chondroepiphysis.
Failure of telescoping and subsequent intramedullary migration of the rod
have been reported in several clinical
series6,7,9-11,13-16,20.
A telescopic rod elongates only when the forces generated by proximal and
distal fixation points exceed the friction forces between the sleeve and the
obturator and between the hardware and the surrounding tissue. However, there
appears to be a gray zone since, although the rod appears straight and
continues to telescope to some extent, the fixation end of the device may cut
through the physis (Fig. 5).
Karbowski et al. reported that the tibia has a higher rate of this
complication than does the femur and that cutting through by the T-piece
occurs more frequently at the distal part of the
tibia16. We believe
that the interlocking pin used with the interlocking telescopic rod, if
properly implanted, provides stronger anchorage than does a T-piece. However,
the obturator migrated proximally along with its interlocking pin in four of
our thirty-two cases. If this complication occurs, we recommend no additional
surgery; instead, we believe that the implant should be allowed to function as
a non-elongating rod. Revision surgery is indicated only if substantial
angulation, symptomatic anterior cortical cutting through, or a fracture
developed at the tip of the outgrown rod.
All patients in the present series received bisphosphonate therapy for
varying periods to strengthen the cortical bone and thereby prevent gradual
angulation of the diaphysis, which helped to minimize the chance of the
telescoping failing. However, the bisphosphonate may not have contributed
appreciably to the strength of the distal fixation because additional bone
formation on the epiphyseal side is less apparent than on the metaphyseal
side21.
It is difficult to compare the survival of the interlocking telescopic rod
with that of previously described telescopic rods because all series differ in
terms of the definitions of complications or failure, surgical procedures used
for insertion, and patient profiles. Marafioti and
Westin3 reported
that the three-year rate of survival of the elongating rods in their series
was 77%. In our series, all of the rods survived for three years
postoperatively, and the four-year survival rate was 89%.
In summary, we described a new telescopic intramedullary rod system that
does not cause injury to the distal articular cartilage of the long bones and
that anchors the obturator to the distal epiphysis. The interlocking
telescopic rod also appears to be easier to remove, with less trauma, during
revision surgery. Our interim results show survivorship comparable with, or
better than, that reported for a telescopic rod with a T-piece. Long-term
follow-up until skeletal maturity or longer will be necessary to fully
evaluate the benefits and limitations of this system. ?