Intrasubstance tears of the posterior cruciate ligament and posterolateral
structures of the knee, a relatively common injury in adults, are very rare in
children and adolescents. Most injuries of the posterior cruciate ligament in
this age-group are osteochondral avulsions of either the femoral or tibial
attachment1-12.
When nonoperative treatment or primary repair of a torn or avulsed posterior
cruciate ligament fails and a child or adolescent experiences instability,
meniscal damage, and early degenerative changes, the physician is confronted
with a dilemma: continued nonoperative treatment will probably result in
progressive deterioration of the knee, but surgical intervention may cause an
iatrogenic physeal injury.
We report the case of an adolescent patient with posterior instability and
posterolateral instability of the knee for whom nonoperative treatment had
failed and who was successfully treated with use of a physeal sparing
intra-articular reconstruction of the posterior cruciate ligament and an
extra-articular reconstruction of the posterolateral structures.
The patient's family was informed that data concerning the case would be
submitted for publication.
The patient sustained a hyperextension injury of the right knee while
jumping on a trampoline when he was eight years old. He was initially seen by
a local physician who made a diagnosis of a sprained knee. The pain and
swelling of the knee gradually subsided without treatment. He sustained a
second injury of the right knee while playing football when he was eleven
years and eight months old. Magnetic resonance imaging revealed partial tears
of the anterior and posterior cruciate ligaments and a medial meniscal tear.
Arthroscopy, which was performed at another institution, revealed a small
radial tear in the white zone of the medial meniscus with an unstable
component, and approximately 10% of the meniscus was excised. The articular
surfaces of the medial femoral condyle and the medial tibial plateau were
normal. Despite its appearance on magnetic resonance imaging, the anterior
cruciate ligament appeared normal at arthroscopy. The patient returned to
sports, although he continued to have mild aching in the medial aspect of the
knee. One and a half years later, he injured the right knee for the third time
while playing football. He was referred to our office when he was thirteen
years and four months old. At that time, he complained of giving-way,
swelling, and pain. Physical examination of the knee revealed a range of
motion of 0° to 135°, a grade-2+ effusion, grade-3+ total
anterior-posterior laxity at 30° of flexion, grade-2+ to 3+ posterior
laxity at 90°, and grade-3+ posterior-lateral
laxity13. The
pivot-shift test was negative, and the external rotation test in the prone
position was >20° for the injured knee compared with the uninjured knee
at both 30° and 90° of flexion.
Examination with a KT-1000 arthrometer (MEDmetric, San Diego, California)
at 134 N (30 lb) of force revealed a side-to-side difference in total
anterior-posterior laxity of 7 mm at 30° and 13 mm at 70° of flexion.
Stress radiography with the Telos device (Austin and Associates, Fallston,
Maryland) performed at 90° of flexion with 89 N (20 lb) of posterior
force14 revealed a
side-to-side difference of 15 mm of posterior displacement of the tibia on the
femur of the involved knee compared with the uninjured knee. Anteroposterior
and lateral radiographs of the knee were normal and demonstrated open physes
of the distal aspect of the femur and proximal part of the tibia. Long-leg
radiographs revealed no limb-length discrepancy or angular deformity. The
score on the 2000 IKDC (International Knee Documentation Committee) knee
examination13 was D
(severely abnormal), and the score on the 2000 IKDC subjective knee
evaluation15 was 47
of 100. Examination of secondary sexual characteristics indicated that the
patient was at Tanner stage
III16 of
maturation. Radiographs of the hand of the
patient17 revealed
that his bone age was thirteen years and eight months.
Arthroscopic examination, performed when the patient was thirteen years and
five months old, revealed a prior partial medial meniscectomy with diffuse
Grade-I changes of the articular cartilage of the medial femoral
condyle18. The
anterior cruciate ligament was intact, and there was a large bucket-handle
tear of the lateral meniscus within 3 mm of the synovium. The lateral meniscus
was repaired. A physeal sparing reconstruction of the posterior cruciate
ligament and reconstruction of the posterolateral structures was also
performed to stabilize the knee, improve the success rate of the meniscal
repair, and delay the progression of chondral degeneration in the medial
compartment.
Surgical Technique
The injured limb was placed in an arthroscopic leg holder with the hip
flexed 20° to facilitate c-arm fluoroscopic visualization of the knee in
the lateral plane. The c-arm was positioned on the side of the table opposite
the injured knee, and the monitor was placed at the head of the table. The
tibial and femoral growth plates were visualized in both the anteroposterior
and lateral planes.
The arthroscope was inserted through the anterolateral portal, and an
anteromedial portal was established. After examination of the joint, the
tibial drill-guide was inserted through the anteromedial portal; however, the
point of the guide was not long enough to insert a tibial pin that would exit
distal to the posterior tibial physis. Consequently, an anteromedial
arthrotomy was performed. The semitendinosus and gracilis tendons were
dissected free through the lower end of the incision, transected at the
musculotendinous junction with the use of tendon stripper, and detached
distally. A number-2 FiberWire suture (Arthrex, Naples, Florida) was placed in
each end of both tendons with an interlocking whip-stitch. The double tendons
were then placed under 4.5 kg (10 lb) of tension on the back table with the
use of a GRAFTMASTER device (Acufex; Smith and Nephew, Andover,
Massachusetts). A lateral incision was then made for the posterolateral
ligament reconstruction. The lateral meniscus was repaired with an inside-out
technique with use of 2-0 Ethibond sutures (Ethicon, Somerville, New Jersey)
and use of vertical stitches on the proximal and distal surfaces of the
meniscal tear. The posterior cruciate ligament was examined and was found to
be torn from the femur. The residual scar tissue at the insertion of the
ligament on the femur was incised, and a whip-stitch with number-5 FiberWire
suture was placed in both the anterolateral and posteromedial bundles of the
native posterior cruciate ligament.
A tibial guide was used to place a guide-pin that was inserted from
anterior to posterior and exited just distal to the tibial insertion of the
posterior cruciate ligament and the posterior tibial physis. Insertion of this
pin was performed with c-arm visualization in the lateral plane. The posterior
tibial cortex was penetrated carefully under real-time c-arm observation. The
tibial guidewire was left in place, and two additional guidewires were
inserted with the use of a drill-guide through the medial femoral condyle,
entering the joint in the footprint of the anterolateral and posteromedial
bundles of the posterior cruciate ligament on the femur. Tendon sizers were
then used to measure the diameter of both the double semitendinosus and
gracilis tendon grafts. A 9-mm drill-hole was made in the tibia, with the
index finger of the left hand placed through the lateral incision to protect
the posterior tibial artery. A 7-mm hole was made for the anterolateral
bundle, and a 6-mm hole was made for the posteromedial bundle. The holes were
chamfered to reduce graft abrasion. The length of the tibial hole was
measured, and an Endobutton continuous loop (Smith and Nephew) was chosen so
that approximately 2.5 cm (the length of the interference screw) of the
quadruple hamstring tendon graft remained within the tibia. The tendons were
then placed around the Endobutton continuous loop, and the Endobutton was
placed through a 12-mm Endobutton washer. The tendons were passed from front
to back through the tibia; the doubled gracilis tendon was passed through the
posteromedial drill-hole, and the doubled semitendinosus tendon was passed
through the anterolateral hole in the femur
(Fig. 1). Number-5 FiberWire
sutures in the anterolateral and posteromedial bundles of the native posterior
cruciate ligament were then passed through the respective drill-holes. The
semitendinosus tendon graft, used to replace the anterolateral bundle, was
secured to the femur with a 7-mm BioRCI screw (Smith and Nephew, Memphis,
Tennessee) with the knee in 90° of flexion. The number-2 FiberWire sutures
in the ends of the semitendinosus tendon were also tied around a screw and
post. With the knee in 30° of flexion, the gracilis tendon graft, which is
used to replace the posteromedial bundle, was secured to the femur with a 6-mm
BioRCI screw and sutures in the ends of the gracilis tendon graft were tied
over the screw and post (Fig.
2).
The reconstruction of the posterolateral structures was then performed. The
iliotibial band was split longitudinally. The popliteus tendon was found to be
avulsed from the lateral femoral condyle. A number-2 FiberWire suture was
placed in the end of the popliteus tendon with a whip-stitch. An
anterior-to-posterior drill-hole was made through the fibular epiphysis with
the use of c-arm visualization. Then the periosteum was roughened just distal
to the lateral epicondyle. A fresh tibialis anterior tendon allograft was
passed through the fibular drill-hole and, along with the popliteus tendon,
was secured to the lateral femoral condyle with a screw and spiked washer
(Figs. 2 and
3). The fascia, subcutaneous
tissue, and skin were closed in a routine fashion, and a hinged brace was
applied. Anteroposterior and lateral radiographs made six months after surgery
showed the position of the fixation devices and the tibial drill-hole (Figs.
4-A and 4-B).
Postoperative Rehabilitation
Phase I of rehabilitation was started as soon as the patient awakened from
surgery. He was encouraged to perform quadriceps muscle contractions and
straight-leg raises. Continuous passive motion from 0° to 60° was used
eight hours a day. Range-of-motion and extension exercises in the prone
position were started the day after surgery. He walked with crutches using
toe-touch weight-bearing for six weeks.
Phase II of rehabilitation took place in the second to the eleventh
postoperative weeks. Active knee-extension exercises were begun, along with
patellar mobilization and electrical stimulation of the muscles. The patient
was fitted with a posterior cruciate ligament functional brace six weeks after
surgery. Exercises were introduced into the rehabilitation program in order of
increasing difficulty, including quadriceps muscle-strengthening,
proprioception exercises, functional strengthening, and strengthening
exercises while in a pool. Hamstring muscle-strengthening exercises were not
performed during the first eight weeks.
Phase III of rehabilitation took place in the twelfth to the twentieth
postoperative weeks. This phase included functional strengthening,
straight-line jogging, plyometric exercises, sport-cord exercises for jogging,
lateral movement, and foot agility exercises. At twenty-four weeks, he was
permitted to perform functional activity, including full-speed running, while
wearing the brace. He was able to advance to full activity, including
competitive sports, thirty-six weeks after surgery.
Postoperative Evaluation
The patient was reexamined when he was fifteen years and ten months of age,
twenty-nine months after surgery. The evaluation included subjective
assessment and objective examination.
According to the criteria of the IKDC subjective form, he was able to
participate in very strenuous activity without pain, swelling, or instability,
with an improvement in the score from 47 preoperatively to 91 at the time of
follow-up. With use of a 10-point scale (with 10 indicating normal, excellent
function), he rated the function of the involved knee as a 10.
Examination revealed full knee extension and a 3° loss of flexion.
There was no effusion, tenderness, or crepitation. Ligament laxity testing
with the KT-1000 arthrometer at 134 N (30 lb) revealed a decrease in mean
side-to-side total anterior-posterior laxity from 7 mm preoperatively to 1 mm
postoperatively at 30° of flexion and from 13 mm preoperatively to 3.5 mm
postoperatively at 70° of flexion. Stress testing with the Telos device at
89 N (20 lb) revealed a decrease in the side-to-side posterior tibial
displacement from 15 mm preoperatively to 4.5 mm postoperatively at 90° of
flexion. The external rotation test in the prone positiondecreased from
>20° at both 30° and 90° of flexion to no demonstrable
side-to-side difference. The overall growth of the patient from the time of
surgery to follow-up was 10 cm. Teleroentgenograms performed with a
radiographic ruler between the legs revealed no difference in limb lengths,
and the tibiofemoral valgus angle was equal in both limbs. Finally, the
objective IKDC knee examination score improved from severely abnormal prior to
surgery to nearly normal at the time of the latest follow-up.
Despite an increased recognition of anterior cruciate ligament tears in
children and adolescents, reports of posterior cruciate ligament tears remain
rare in skeletally immature patients. Our review of the literature revealed
only fourteen
avulsions1,3-12
and one intrasubstance tear of the posterior cruciate ligament in this
age-group19. The
difference in the prevalence of such injuries is due to maturation-related
anatomical and biomechanical factors that predispose skeletally immature knees
to physeal and bone injury rather than to posterior cruciate ligament tears.
Tears rarely occur in the posterior cruciate ligaments of skeletally immature
patients because the ligament and dense attachment of the perichondrium are
stronger than the hypertrophic zone of the physis. Consequently, almost all of
the reports of posterior cruciate ligament injuries in the age-group have been
avulsions from the femur.
The natural history of posterior cruciate ligament tears in children and
adolescents has not been documented, although it can be extrapolated from the
case reports of pediatric patients and the results of studies published on
nonoperative treatment in adults. Most investigations of adult
patients6,20-23
have documented good subjective outcomes with nonoperative treatment of
isolated posterior cruciate ligament
tears23-27.
However, some have concluded that residual posterior laxity caused by isolated
tears is not totally
benign22,28-32.
Alterations in the kinematics of the posterior cruciate ligament-deficient
knee result in elevated contact pressures that can cause later degenerative
changes of the patellofemoral and medial compartments of the
knee33,34.
The prognosis for knees in which there is a posterior cruciate ligament tear
combined with other ligamentous injuries is generally poor when the knees are
treated
nonoperatively23.
The results in the few reports of nonoperative treatment of posterior cruciate
ligament injuries in children and adolescents appear to be similar to those
reported for adults. The results are good in some
patients3,19
and poor in
others2,12.
Our approach to the management of posterior cruciate ligament tears in
adults consists of nonoperative treatment of isolated tears and surgical
reconstruction for tears combined with other ligament injuries. Reconstruction
was considered in this patient because nonoperative treatment had already
failed and the patient had instability, a lateral meniscal tear, and
degenerative changes in the medial joint space.
Most surgeons have recommended repair of posterior cruciate ligament
avulsions in this age-group, and the results have been generally
good4,6,8,9,11,12.
Even so, the repair may fail and, in that case, nonoperative management is
indicated. When a child or an adolescent has persistent symptoms despite
nonoperative treatment, should surgical reconstruction in a child be
performed? Frank and
Strother3 stated
that posterior cruciate ligament reconstruction in a child is without
documented precedent. Our review confirmed this observation.
The primary reason for concern about reconstructing posterior cruciate
ligaments in adolescents is the potential consequence of iatrogenic physeal
injury. Some adolescent patients have a great deal of growth remaining, while
others have minimal growth of the distal femoral and the proximal tibial
physis. The relative surgical risk of growth disturbance can be estimated by
considering the chronologic age of the patient, the skeletal age as determined
by radiographs, and the physiologic age as established by the Tanner staging
system of sexual maturation. Studies of normal growth can be used to estimate
the consequences of limb-length discrepancy.
Pritchett35
reported that the distal part of the femur grows 1.3 cm per year until the
last two years of maturity, when the growth rate decreases to 0.65 cm per
year. The rates of the proximal tibial growth are 0.9 cm per year and, in the
last two years, 0.5 cm per year. The average boy at thirteen years and six
months old would then have just over 4 cm of growth remaining in the distal
femoral and proximal tibial physes. Although there was a potential for
substantial limb-length discrepancy in the patient in the present report, this
risk was thought to be relatively small because the physeal sparing
drill-holes were unlikely to cause a complete arrest of both the tibial and
femoral physes. Therefore, our greatest concern was not length discrepancy but
angular deformity. Wester et
al.36 estimated
that, in the worst-case scenario, a fourteen-year-old boy with 2 cm of
remaining distal femoral growth could have development of a 14° angular
deformity with femoral epiphysiodesis or an 11° angular deficiency with
partial arrest of the proximal tibial physis. The risk of angular deformity
was minimized in our patient by the use of a soft-tissue graft, creation of
all epiphyseal femoral and fibular tunnels, placement of the tibial tunnel
exit inferior to the tibial physis, and avoidance of soft-tissue dissection
near the perichondrial ring of either the tibial or femoral physes.
Research on the effects of physeal injury in animal models and our prior
clinical experience with transepiphyseal reconstruction of the anterior
cruciate ligament were both considered when this operation was
planned37,38.
In general, the risk of growth disturbance is determined by the extent of
damage to the cross-sectional area of the
physis39-42.
However, there is still uncertainty—even in animal
models—regarding the safety of tensioning the graft around the physis
and the size of transepiphyseal drill-holes that can be created without
causing growth
disturbance37.
Although placement of a soft-tissue graft across the physis appears to offer
protection from growth
arrest43, the
results of animal studies have been
contradictory40,41.
Our fear of iatrogenic growth disturbance was allayed by the fact that
transphyseal drill-holes were not performed in this procedure and by our
previous experience that demonstrated the relative safety of the exclusive use
of epiphyseal
drill-holes38,44.
The success of posterior cruciate ligament reconstruction depends on many
factors, including graft selection, tunnel position, and methods of fixation.
Our preferred technique of reconstruction in adults is to use a tibial inlay,
double-bundle bone-quadriceps tendon graft. The unique challenge in the
patient in the present report precluded the use of this technique. Autologous
semitendinosus and gracilis tendons were used instead of the quadriceps tendon
because of the distinct advantages of the use of hamstring grafts in
prepubescent patients. Quadruple hamstring grafts have high strength and can
be harvested with low donor-site morbidity. They are relatively long and are
composed entirely of soft tissue. If the physis is inadvertently broached
during surgery, filling the hole with soft tissue can reduce the risk of
growth
disturbance41,43.
Tibial inlay, a technique that was developed to eliminate the acute graft
angle that occurs with a tibial
drill-hole45,46,
could not be used in this patient. Inlay of a bone block at the insertion site
on the tibia would damage the posterior aspect of the tibial physis. A
double-bundle reconstruction was performed because biomechanical studies have
demonstrated that it may result in more normal knee laxity over the full range
of knee motion and more normal graft force in the posterior cruciate
ligament47,48.
Recent clinical reports have also suggested that a two-strand graft may
improve the results of
reconstruction48,49.
A disadvantage of this posterior cruciate ligament reconstruction, modified
to accommodate the open physes, was the nonanatomic tibial tunnel placement
distal to the insertion of the posterior cruciate ligament. Tibial tunnel
placement in this position may be associated with residual laxity. The acute
angle that the graft must take around the proximal aspect of the posterior
part of the tibia may also lead to high graft forces and graft abrasion.
Despite these disadvantages, the risk of failure imposed by this technique was
probably less than the risk of iatrogenic growth disturbance from anatomic
transphyseal placement of drill-holes.
Previous studies have demonstrated that if the deficiency of the
posterolateral structures is not corrected, the graft is subjected to high in
situ forcesthat can cause graft
failure50,51.
We found that the fibular and femoral physes could easily be avoided while a
standard posterolateral ligament reconstruction was performed. The
posterolateral ligament reconstruction, like the posterior cruciate ligament
reconstruction, was a physeal sparing procedure. The identification and repair
of the torn popliteus tendon was a major contributor to the success of the
posterolateral ligament
reconstruction50,52.
The majority of posterior cruciate ligament injuries in pediatric patients
are avulsions that can be treated successfully with surgical repair. The case
of our patient demonstrates that reconstruction can be successful in the rare
instance of an adolescent patient who has persistent symptoms after
nonoperative treatment or failure of a posterior cruciate ligament repair.
Although the patient did not have growth disturbance, any procedure that
includes transepiphyseal drilling or dissection adjacent to the physis has the
potential to cause growth disturbance. ?