Lengthening for congenital femoral deficiency (CFD) is most applicable in femurs with an intact mobile hip and knee joint irrespective of the amount of discrepancy. Pre-existing knee stiffness is the most functionally limiting factor and should be considered a relative indication for amputation vs. reconstruction. Hip dysplasia or deficiency is reconstructable and is not a limiting factor. Hip reconstruction should be performed prior to lengthening.
Knee instability of mild degrees can be protected by stabilizing the tibia during the lengthening. Knee or patellar dislocation should be corrected by a modification of the Langenskiold procedure (for congenital patellar dislocation) prior to lengthening. Delayed ossification of the femoral neck should also ossify prior to lengthening. It may be necessary to perform a valgus osteotomy to eliminate shear in the femoral neck and allow earlier ossification of the proximal femur. Coxa vara can be corrected at the time
of lengthening but neck shaft angles less than 90 deg should be corrected prior to lengthening. Pelvic osteotomy to correct hip dysplasia when the center edge angle is less than 20 deg should also be performed prior to lengthening. multiple lengthenings are required for most discrepancies. These should be spaced out as much as possible during childhood. Most lengthenings are completed before high school. The maximum lengthening should not exceed 8-l2cm. in older children and 4-6cm in younger children. Soft tissue releases of the fascia lata, rectus, hamstring and adductor tendons is recommended
with each lengthening. Botox injection is another helpful adjunct. Intensive physical therapy to maintain motion of the knee and hip is an essential component of the lengthening treatment. Limiting factors to reconstruction are psycho-socio-economic factors and the lack of experience by the treating orthopedic surgeon.
In our experience good or excellent results are obtainable in 91% of children for this condition despite a 70% complication rate with 50% of cases requiring additional surgery to treat complications. Most complications do not lead to permanent sequelae that would worsen the functional result if they are treated aggressively.
The results of lengthening for CFD were the same as for traumatic and developmental limb length discrepancy in a group of 70 femoral lengthenings performed by the author.
The congenital femoral deficiency (CFD) is a spectrum of severity of femoral
deficiencies and deformities. Deficiency refers to lack of integrity, stability
and mobility of the hip and knee joints. Deformity refers to mal-orientation
and malrotation and contractures of the hip and knee. Both deficiencies and
deformities are present at birth, non progressive and of variable degree.
Classifications of congenital short femur and proximal femoral focal deficiency
are descriptive but are not helpful in determining treatment. A recent
longitudinal follow-up of different classification systems' showed that they
were inaccurate in predicting the final femoral morphology based on the initial
radiograph. My classification system is based on the factors that influence
the lengthening and reconstruction of the congenital short femur.
Congenital Femur Deficiency Classification(Paley)
Type 1: intact femur with mobile hip and knee
a) normal ossification proximal femur
b) delayed ossification proximal femur
Type 2: mobile pseudarthrosis with mobile knee
a) femoral head mobile in acetabulum
b) femoral head absent or stiff in acetabulum
Type 3: diaphyseal deficiency of femur
a) knee motion > 45 degrees
b) knee motion < 45 degrees
Knee joint mobility/deficiency and not hip joint mobility/deficiency is the
most determinant factor for functional outcome and reconstructability of
the CFD (e.g.- the degree of hip deficiency does not change following
amputation-yet increasing degrees of hip deficiency is used as an indication
for amputation and prosthetic fitting). Therefore Types 1 and 2 are the most
reconstructable. A wide spectrum of hip and knee dysplasia/deficiency and
deformity exists in Type 1 cases. Since this is the type most amenable and
most commonly lengthened it is worthy of subclassification according to factors
that require correction before the lengthening can be carried out. These
factors affect the age at which the lengthening process can begin since multiple
corrections will delay the first lengthening. They also affect the number
of surgeries that are required prior to starting the lengthening and therefore
may affect the decision of reconstruction vs amputation or rotationplasty.
Type 1 Subclassification
0) ready for surgery; no factors to correct before lengthening
1) one factor to correct before lengthening
2) two factors to correct before lengthening
3) three factors to correct before lengthening
(factors requiring correction prior to lengthening of femur: NSA < 90
deg +/- delayed ossification proximal femur, CE angle < 20 deg, subluxing
patella and/or dislocating knee)
NSA < 90 deg is corrected by proximal femoral osteotomy (this is also
the treatment if there is delayed ossification of the proximal femur); CE
angle < 20 deg is corrected by pelvic osteotomy; subluxing patella and/or
dislocating knee are corrected by modifications of the Langenskiold procedure
The strategy of management of all
Type 1 cases is to convert Type lb into Type la, and Types la-l,
la-2, la-3, etc into la-0. Type la-0 cases can be treated by limb length
equalization by one or more lengthenings and/or epiphyseodesis or femoral
The strategy of management for
Type2 cases is to first determine there is either a mobile pseudarthrosis
or an absent femoral head. It is important not to mistake Type lb cases for
Type 2 and vice versa. In Type 2a cases where there is mobility between the
femoral head and the acetabulum and between the femoral shaft and the femoral
head, union of the pseudarthrosis is an initial goal of treatment converting
Type 2a to Type la. On the other hand Type 2b cases where there is a femoral
head with no mobility in the acetabulum cannot be converted to Type lb since
the hip is stiff. In Type 2b cases with or without a femoral head a pelvic
support hip reconstruction procedure is combined with one of the lengthenings.
Until then lengthening can be carried out with extension of the lengthening
device to the pelvis. Type '2 cases can be treated by limb length equalization
by one or more lengthenings and epiphyseodesis or femoral shortening procedures.
The strategy of management for
Type3 cases is to first determine the range of knee joint motion.
If the knee has less than 45 deg of motion and the ankle has a good range
of motion then a Van Nes rotationplasty with prosthetic fitting should be
considered. If the ankle has a poor range of motion in then a Symes amputation
combined with prosthetic fitting is the preferred option. Both of these
prosthetic reconstruction options can be combined with a pelvic support osteotomy
at a latex date. Equalization of limb length by multiple lengthenings is
possible but yields the functional outcome of an arthrodesed knee; 'a long
run for a short slide'. If there is a significant diaphyseal proximal deficiency
but there is more than 45 deg of knee motion present then lengthening can
be considered with all efforts made to preserve the knee joint function.
The goal of lengthening reconstruction surgery in these cases is to convert
the type 3a femur to a type 2b femur. The rest of the reconstruction is as
for type 2b. This type of reconstruction while feasible is very complicated
and prosthetic reconstruction options may be preferable to minimize the
complexity of treatment.
Based on the clinical exam and the initial radiograph one should try to determine
if the femur has a femoral head or a pseudarthrosis. If the femoral head
is present without a mobile pseudarthrosis then the femur is
considered intact. In the intact femur the neck shaft angle should be
evaluated for varus and the acetabulum for dysplasia. If a pseudarthrosis
line is present, the pseudarthrosis should be examined fluroscopically to
determine if it is mobile or stiff. In a stiff pseudarthrosis the femoral
head moves with the rest of the femur. In a mobile pseudarthrosis the femoral
head completely or partially doesn't move with the rest of the femur. The
unossified femoral neck gives the appearance of a pseudarthrosis. In the
infant it is necessary to perform an arthrogram under anesthesia to determine
if there is an intact femoral neck or not and whether a pseudarthrosis is
stiff or mobile. If the pseudarthrosis is mobile then it is important to
establish if the femoral head is at all mobile within the acetabulum.
Abduction-adduction, flexion-extension and internal-external rotation movements
under image intensification with the dye in the joint will demonstrate if
the femoral head moves in the acetabulum. Push-pull stress movements with
the dye in the joint will demonstrate if a mobile pseudarthrosis is present.
In cases where no motion of the femoral head is appreciated the spinal needle
used for arthrography can be inserted. into the femoral head and manipulated
from the outside under image intensifier observation. This test may be difficult
to assess. The presence of a well developed acetabulum is the best clue that
the femoral head is present and probably mobile. MRI and ultrasound
may be helpful but are often difficult to interpret in infants with such
abnormal anatomy. The arthrogram performed by the orthopedic surgeon remains
my preferred testing modality
It is important to examine the range of motion of the hip and identify the
presence of fixed adductus, flexion and external rotation of the hip. Fixed
flexion deformity of the hip should be compared to the other side in infants
before walking age. There is normally some FFD in infant hips for many months
after birth. Lack of abduction is a sign of varus deformity of the hip rather
that dislocation of the hip. External rotation of the limb is also typical
of all grades of CFD.
The initial evaluation of the knee is also performed clinically and
radiographically. The range of motion of the knee is a critical feature.
FFD of the knee is normally present in infant knees. Therefore the affected
side should be compared to the normal side. If there is more FFD on the affected
side. this is significant. A stiff knee is rarely found except in patients
with significant proximal femoral deficiency. In these patients it may be
difficult to assess the ROM of the knee because of the short chubby thigh.
Examination under anesthesia may be required together with an arthrogram.
In stiff knees the plain radiograph may suggest flattening of the posterior
half of the femoral condyles as the cause of the FFD. This can be confirmed
by arthrogram. The arthrogram may also show a lack of a patello-femoral synovial
pouch and an absent patella.
In patients with a mobile knee, the knee should also be evaluated for stability
of the tibio-femoral joint and for the presence and tracking of the patella.
The stability of the knee in the AP and ML directions should be noted as
well as torsional stability. The patella should be palpated and its tracking
observed. It may be absent, dislocated, dislocating, subluxing or stable.
The tibia may dislocate as it goes into full extension due to tight lateral
structures (fascia lata and hamstrings), incompetent capsulo-ligamentous
structures and anterior deficiency of the femoral condyles. The knee flexion
angle of dislocation/relocation should be noted as well as the direction
of rotation of the tibia on the femur with dislocation and reduction.
As noted above this group is the most reconstructable. Lengthening treatment
in these patients should not begin until the neck shaft angle is greater
than 90, CE angle > 20, a nonsubluxing patella and nondislocating knee.
Each of these prerequisites will be examined separately.
If the neck is ossified but the NSA is less than 90 deg then a proximal femoral
valgus subtrochanteric osteotomy should be performed prior to lengthening.
fly preference is to perform this with external fixation. The hip osteotomy
should correct the varus, the bony FFD, and the external rotation deformity.
If the hip osteotomy is performed with internal fixation then the hardware
needs to be removed before the lengthening. To avoid this additional procedure,
reduce the incisional scar of a plating procedure, and to increase the accuracy
of correction of a complex valgus, lateral translation, extension and external
rotation osteotomy I prefer to use the Ilizarov device to perform the hip
osteotomy. External fixation also seems less limited in the amount of correction
it can achieve.
For neck shaft angles less than 90 deg especially in infants and young children
I prefer to perform the hip osteotomy separate from the lengthening procedure
(fig 2). Technically both can be performed at the same time, however, large
valgus corrections of the proximal femur apply considerable pressure to the
hip joint from the large acute lengthening that occurs. Lengthening of the
femur at the same time would add additional pressure to the hip joint. We
therefore prefer to separate the two procedures. Smaller degrees of coxa
vara stabilize the slightly dysplastic acetabulum and should not be corrected
prior to lengthening(fig 3). NSA between 90-110 deg should be considered
for correction together with lengthening especially if the CE angle is greater
than 25 deg or if the femoral head physis is vertically inclined. (in the
latter situation there is a risk of physeal slip or separation during
Technique for hip osteotomy: A percutaneous
adductor tenotomy is performed first. Two half pins are inserted by the canulated
drill technique into the proximal femur while holding the hip maximally adducted
and externally rotated(placing the hip in its neutral position). These pins
are parallel to the line from the tip of the greater trochanter to the center
of the femoral head. Two half pins are inserted into the diaphysis and distal
femur using the canulated drill hole technique with the limb in a neutral
position. These pins are parallel to the knee joint line and to each other.
One Ilizarov arch or segment of a half ring is connected to each pair of
pins perpendicular to its sepnent of bone. A percutaneous multiple drill
hole osteotomy is performed perpendicular to the femoral diaphysis just distal
to the lesser trochanter. The bone is first internally rotated, then laterally
translated, then angulated into valgus and tonally extended. The order of
correction is critical to maintain bony contact. The arches are fixed to
each other using threaded rods with conical washers to hold the correction.
If the osteotomy is performed for a delayed ossification of the femoral head
a smooth K-wire or small half pin can be inserted across the femoral neck
to prevent fracture from the valgus stress of the correction. There will
be a fixed abduction contracture at the end of the valgus correction. This
will stretch out spontaneously over time.
If the CE angle is less than 20 deg then a pelvic osteotomy should be performed
to stabilize the hip before a lengthening procedure (2). If a proximal femoral
osteotomy is necessary to correct a severe hip varus, it should be performed
at a separate time prior to the pelvic osteotomy. A 3D reconstruction CAT
scan of both hips is useful in deciding upon which pelvic osteotomy to perform.
If the 3D scan shows good coverage of the affected hip posteriorly and deficiency
anterolaterally as compared to the other side then a Salter osteotomy may
be performed for coverage. To add some length the Millis-Hall modification
of the Salter osteotomy for pelvic lengthening is performed (3)(fig 4). If
the 3D scan shows decreased coverage posteriorly due to a hypoplastic posterior
lip of the acetabulum then the Salter osteotomy should not be performed since
it will decrease the posterior coverage further. In these cases my preference
is a Dega osteotomy with shelf augmentation if necessary(fig 5).
Dislocation of Patella or Tibia
Dislocation of the patella or tibia with flexion or extension respectively
necessitate a stabilizing procedure prior to lengthening. Isolated instability
of the tibia-femoral joint without dislocation does not need to be addressed
before lengthening. Isolated subluxation of the patella should also be treated
prior to lengthening. I use a modification of the Langenskiold procedure
(procedure designed for congenital dislocation of the patella.) to stabilize
the knee (4). This procedure may be performed at the same sitting as the
pelvic osteotomy since both need to be in a long leg cast postoperatively.
Modified Langenskiold technique: the knee
is exposed through a long S shaped incision. The biceps tendon and fascia
lata is z lengthened if needed. The capsule is separated from the patella
and from the synovium laterally and medially. The synovium is cut from the
patella circumferentially. The quadriceps tendon is left attached to the
patella proximally and the patellar tendon remains attached to the patella
distally. A more medial synovial incision is made for reinsertion of the
patella. The synovium is sewn to the patella circumferentially thus centralizing
the patella. The synovium is closed laterally where the patella was removed.
The patellar tendon is sharply elevated off the apophyseal cartilage and
is left connected to periosteum distally. The patella can be moved medially
as much as needed pivoting on the distal periosteum. To hold it medial the
medial side of the tendon is stitched to medially periosteum in the Grammont
technique (5). The capsule is stitched overtop the pateila on the medial
side and left open laterally.
To stabilize the tibia on the femur the fascia lata and hamstrings are
lengthened. If the tibia is external rotationally unstable to the femur then
the distal attachment of the fascia lata is brought anterior to the patella
and anchored to the medial epicondyle. This is only applicable for
postero-lateral instability of the tibia on the femur(fig 6). After realignment
if the vastus lateralis pull is still a deforming lateral force on the patella
then its distal tendonous portion can be detached and pulled over top the
patella as an additional medial restraint. ) If the dislocation is antero-lateral
then we use the fascia lata looped over itself passing under the lateral
capsule or the lateral collateral ligament to reattach to Gerdie's tubercle
(MacIntosh-lateral substitution-(fig 7) (6). After wound closure the leg
is put in a cylinder cast for 6 weeks followed by passive and active motion.
The above problems must all be addressed before beginning the femoral
lengthening. Once they are corrected the femur is considered a Type la-0
which is ready for lengthening.
The natural history of the intact, unossified femoral neck is probably to
eventually ossify. Radiographically the lack of ossification is often interpreted
as a pseudarthrosis. Arthrographic examination reveals that the neck and
shaft move as one. Whether some of these go on to pseudarthrosis is debatable.
The coxa vara associated with these unossified femoral necks probably contributes
to the shear forces on the neck that delay its ossification. Therefore the
treatment of the delayed ossification of the femoral neck is a valgus proximal
femoral osteotomy. This is performed in the manner described above for the
ossified proximal femur.
One particular type of delayed ossification is a stiff nonunion line in the
inter or subtrochanteric region. There is no movement visible of this nonunion
line even with stress and radiographic examination. This type of nonunion
is always associated with coxa vara. It can be ignored and the rest of the
treatment carried out as for the delayed ossification cases with coxa vara.
One alternative treatment I have used in these cases is distraction of the
pseudarthrosis to correct deformity and lengthen (fig 11 )
Lengthening of Type 1 CFD
Choice of Osteotomy Level for Lengthening of the Congenital Short
The level of the lengthening osteotomy is dependent on the technique used;
external fixator lengthening or lengthening over a nail (external fixator
and nail combined). With external fixation lengthening without a nail the
external fixator remains on the limb for the distraction and the consolidation
phase. Once the fixator is removed the bone is at risk of fracture through
the regenerate or the pin sites. The retained tension in the soft tissues
particularly the adductors and the hamstrings increases this risk of fracture.
These forces are even greater the more proximal the osteotomy site. The adductors
have little effect on distal osteotomy sites but a strong effect on proximal
osteotomy sites. The larger the bone cross section the less likely it is
to bend or break.. Distal osteotomies have larger cross sectional diameters
than proximal osteotomies.
Bone regeneration potential of distal osteotomies is much different than
proximal osteotomies in the CSF. The proximal diaphyseal bone is often sclerotic
with a narrowed medullary canal. The subtrochanteric bone has a poor ability
to regenerate compared to normal proximal femoral bone and compared to the
ipsilateral distal femur. There is also a higher incidence of refracture
with proximal level osteotomies. Therefore lengthening should be avoided
in the proximal femur unless a nail is used at the same time.
Proximal femoral lengthening has less effect on knee ROM. Distal osteotomies
have a greater risk of knee stiffness and subluxation. Proximal osteotomies
have a greater risk of hip subluxation and tend to go into varus with
In addition to these factors one must consider the derotation and varus
correction of the femur which are performed proximally and the valgus correction
of the knee which is performed distally. If both need to be performed then
a proximal osteotomy is performed for derotation and varus, and a distal
osteotomy for the valgus and lengthening(fig 3). If the derotation does not
need to be performed then only the distal osteotomy is performed for the
lengthening and valgus correction. If the femur is completely straight with
perhaps only some rotational deformity the osteotomy can be made in the
mid-diaphysis which has a wider cross sectional area than the proximal femur
and is not the zone of sclerotic poorly healing bone (fig 9).
In older children with a wider medullary canal (greater than 7mm) lengthening
over a nail can be performed(fig 9). In LON a proximal osteotomy can be used
for lengthening since there is little risk of refracture with a rod in the
medullary canal. Intramedullary nailing in children adds the risk of disturbance
of growth of the apophysis (7) and avascular necrosis of the femoral head
(8). To avoid the latter we use a greater trochanteric starting point and
a nail with a proximal bend (e.g. humeral or tibial). To avoid a coxa valga
deformity we prefer to use this technique in patients with some coxa vara.
The apophyseodesis created by the nail can lead to gradual correction of
the coxa vara. Fixator only lengthening is the method we usually use for
the first lengthening. LON is usually the method we chose for the second
lengthening if the anatomic dimensions and deformities mentioned above permit.
External Rotation Deformity:
Almost every congenital short femur has an external rotation deformity(hip
retroversion). This must never be corrected through a distal femoral osteotomy
for fear of subluxation of the patella (fig 8). The quadriceps muscle is
normally rotated with respect to the knee joint. Therefore if derotation
is performed through a distal femoral osteotomy the knee would rotate medially
while the patella would be pulled laterally. These patients normally have
a hypoplastic patella and a fiat patellar groove on the distal femur. Lateral
subluxation or dislocation of the patella would likely occur. If the derotation
osteotomy is performed proximally (subtrochanteric) the entire quadriceps
mass attached to the shaft of the femur rotates medially decreasing the lateral
pull on the patella.
External rotation deformity is corrected at the sarge time as the hip varus
deformity. If there is a large coxa vara as discussed above then the derotation
is performed at the separate procedure prior to the lengthening. If there
is a mild or no coxa vara then the derotation is performed at the time of
lengthening as an acute correction through a proximal femoral osteotomy.
Distal Femoral Valgus Deformity:
The distal femur is almost always in valgus. This is due to hypoplasia of
the lateral femoral condyle. This is not a growth plate related deformity.
Therefore it is nonprogressive. The center of rotation of angulation of this
valgus deformity is at the level oi the knee joint line. Therefore any osteotomy
to correct the valgus in the supracondylar region needs to angulate into
varus and translate laterally in order to avoid creating a secondary
translational deformity. The amount of angulation is usually between 5-10
deg. This deformity does not need to be corrected before the lengthening
but should be corrected with the lengthening. The fascia lata should be
transected or lengthened at the time of correction to help prevent recurrence,
increased pressure on the lateral compartment of the knee, knee subluxation
during lengthening and loss, of knee motion.
Soft Tissue Releases for Lengthening of the CFD:
Patients with CFD usually have a mild FFD of the hip or a lack of hyperextension
of the hip compared to the other side. A positive Ely test may be present.
This indicates that the rectus femoris is the etiology of the FFD or lack
of HE. FFD may also be present at the knee. Some patients do not have a knee
FFD but have an increased popliteal angle compared to the other side. Lack
of hip abduction is often present. This may be related to coxa vara and/or
to tight hip adductors. Finally the fascia lata can be palpated distally
to be thick and tight.
Soft tissue releases are essential in conjunction with the lengthening to
prevent subluxation and stiffness of knee and hip. In the CFD we always release
the rectus femoris tendon at its origin, and the fascia lata. The rectus
tendon is cut through a 3 cm oblique incision. It is exposed in
the interval between the sartorius and the tensor fascia lata. Care is taken
not to injure the lateral femoral cutaneous nerve. The distal fascia lata
is cut through a longitudinal incision at the level of the proximal pole
of the patella. The incision is located over the intermuscular septum so
that the lateral hamstring can also be exposed. The fascia lata is cut
transversely from its anterior thickening to the intermuscular septum. The
septum is also transected. The posterior extension of the fascia lata is
also transected. Through the same incision the lateral hamstring muscle is
exposed. The tendonous portion of the muscle is cut leaving the underlying
muscle in continuity. If there is a popliteal angle over 10 then a separate
incision is made to fractionally lengthen the semimembranosis, and gracilis
tendons. The semitendinosis tendon is transected. Through the fascia lata
incision the anterior fascia of the thigh which is the anterior extension
of the fascia lata is also transected. The adductor muscles are also released.
For distal femoral lengthenings a percutaneous release is sufficient. For
proximal femoral lengthenings we prefer an open more extensive adductor release.
Instead or as an adjunct to soft tissue release we have recently started
to use Botulinum toxin to temporarily weaken or paralyze some of the hamstrings
and adductors and even the rectus femoris. This is injected at the time of
surgery, at a dose of 5 units/ kg total body dose, with between 30-50 units
into each muscle head. Its effect should wear off in 3-6 months. Botox seems
to reduce muscle spasm and pain in these patients.
Finally the timing of soft tissue release may be an important factor. I usually
perform the release at the time of application of fixator and lengthening
osteotomy. The problem with this is that the soft tissues are not under tension.
They therefore heal up during the lengthening and retether the femur. Recently
I noticed that in a few cases where the soft tissue release was performed
as a planned second stage procedure (4-6 weeks after the lengthening began),
I was able to achieve greater lengthening because the knee motion was better
maintained for longer. Delayed soft tissue release cuts the soft tissues
when they are under tension and prevents them from joining up before the
distraction is over.
Knee instability consideration:
Almost all CSF can be assumed to have hypoplastic or absent cruciate ligaments
with mild to moderate AP instability. Some also have ML and torsional
instability. Despite this the knee tracks normally preoperatively and there
is no indication to do a ligamentous reconstruction in most cases. The
significance of the knee instability to lengthening is the tendency to
subluxation of the knee with lengthening. Knee subluxation with lengthening
is usually posterior, or posterior plus external rotation but can also be
anterior. Posterior subluxation can only occur with knee flexion. Therefore
to prevent posterior subluxation some people recommend splinting the knee
in extension thoughout the distraction phase (9). This promotes knees stiffness
while protecting the knee from subluxation. The preferable way to protect
the knee is to extend the fixation to the tibia with hinges. The hinges permit
knee motion while preventing posterior as well as anterior subluxation. For
this reason we prefer to use a fixator that allows hinging across the knee
joint. This is easily performed with the Ilizarov circular (fig 10a) fixator
but not as readily with the monolateral fixators (fig 10b).
A less common knee instability is anterior dislocation of the tibia on the
femur. This type of dislocation occurs as the knee goes into extension. It
is important to document at which angle of flexion the knee relocates (conversely
at which angle short of full extension the knee dislocates. The dislocation
is due to an anterior deficiency of the distal femur (the lateral radiograph
of the knee shows a lack of the anterior protuberance of the femoral condyles).
One treatment of this instability is extension osteotomy of the knee. The
distal femur is extended by the number of degrees of flexion required to
relocate the knee. I have found the modified Langenskiold procedure better
than extension osteotomy in these instances. Knee extension osteotomy leads
to loss of knee flexion.
(Distal Femoral Lengthening Technique: The
soft tissue releases are performed first. Two Ilizarov rings properly sized
for the distal femur are applied to a distal femoral reference wire. The
angle between the rings is equal to the valgus deformity. The rings are connected
by juxta-articular hinges that are distal to the lower ring. A half pin is
connected to the proximal ring. at the level of the distal third of the femur.
Two additional half pins are inserted at the lesser trochanteric level and
fixed to an arch which is fixed to the proximal ring. Two half pins are inserted
posteromedial and posterolateral at the level of the distal metaphysis and
fixed to the distal ring. The one wire can be removed. Hinges are attached
to the distal femoral ring after identifying the level of the center of rotation
of the knee joint using the image intensifier. the center of rotation of
the knee is located at the virtual intersection of the posterior femoral
cortical line and the distal femoral physeal line (10). Two tibial half pins
are inserted and fixed onto a half ring. This is connected to the hinges.
At least a 90 deg range of knee motion should be possible with the hinge.
The distal femoral osteotomy is performed percutaneously with multiple drill
holes and an osteotome. If a simultaneous derotation and mild coxa vara
correction are also planned then a proximal osteotomy for correction of deformity
along with the distal lengthening osteotomy is needed. Two half pins are
used at each of the three levels of bone. )
Rehabilitation and Follow-up during lengthening:
Femoral lengthening requires close follow-up and intensive rehabilitation
in order to identify problems and maintain a functional extremity respectively.
Follow-up is usually every 2 weeks for radiographic and clinical assessment.
Clinically the patient is assessed for hip and knee range of motion, knee
subluxation, nerve function, and pin sites. Radiographically the distraction
gap length, regenerate bone quality, limb alignment, and joint location are
Knee flexion should be maintained at greater than 45 deg. If knee flexion
is 40 deg or less then the lengthening should be stopped and the knee
rehabilitated more. If after a few days the knee flexion greater than 45
deg is regained, lengthening may resume. Remember never sacrifice function
for length. We can always add more length at an additional lengthening but
we cannot recreate a knee joint. Knee extension should also be monitored.
Maximal extension is usually present following surgery. A flexion contracture
may develop during lengthening. To prevent this a knee extension bar may
be used at night and for one or two hours during the day. A fixed flexion
deformity of the knee places it at risk of posterior subluxation. Subluxation
of the knee can be suspected clinically based on a change in shape of the
front of the tibia relative to the kneecap. Posterior subluxation of the
tibia presents with a very prominent knee cap and a depression of the tibia
relative to the kneecap (skihill sign). Extension of the external fixation
across the knee with hinges prophylaxes against posterior subluxation (11).
Hip motion may become more limited with lengthening. Adduction and flexion
contractures are the most significant since they lead to hip subluxation
and dislocation. Rerelease of the adductors and the rectus, sartorius and
the tensor fascia lata during lengthening may need to be considered in order
to allow continued lengthening.).
The deep peroneal nerve is the nerve at greatest risk with femoral lengtheaing.
Referred pain to the anterior distal leg or dorsum of the foot should be
considered peroneal nerve related until proven otherwise. Hyper or hypoesthesia
in the distribution of the peroneal nerve or weakness of the extensor hallucis
longus muscle are corroborative evidence of nerve entrapment. A nerve conduction
study (our preference is near nerve conduction using very fine needle technique
at the level of the fibular neck) may show evidence of nerve injury. Finally
quantitative sensory testing if available is the most sensitive test to assess
for nerve involvement. With quantitative sensory evidence and sensory signs
only the distraction is slowed to see if the referred pain goes away. If
the referred pain does not dissipate or if motor signs or positive nerve
conduction evidence of nerve injury are present, a nerve decompression at
the neck of the fibula is carried out. Lengthening may continue after the
decompression at 3/4 or 1/2 mm/ day.
Hypotrophic regenerate formation requires slowing down the distraction rate
Overabundant bone formation that may lead to premature consolidation requires
speeding up the distraction rate for few days. A mismatch between the increase
in the distraction gap from one visit to the next and the number of millimeters
of distraction carried out during the same time period is a sign of an impending
premature consolidation. Radiographs are also used to assess joint location.
A break in Shenton's line or increased medial-lateral head-teardrop distance
indicates subluxation of the hip. In the knee posterior or anterior subluxation
can be monitored on the lateral full knee extension radiograph". Limb length
equalization should be based on full length standing radiographs. If there
is a knee flexion deformity a scanogram with the knees equally flexed and
the hip, knee and ankle equally positioned to the radiographic plate is used
instead. Limb alignment is assessed for femur and tibia separately and in
combination. Separately the joint orientation of the knee should be measured
using the Malalignment Test (12). Axial deviation from lengthening (procurvatum
and valgus for distal femoral lengthening and procurvatum and varus for proximal
lengthening) is identified and corrected at the end of the distraction phase
when the regenerate bone is still malleable.
When there is malalignment of the femur and tibia, the femoral malalignment
is corrected to a normal distal femoral joint orientation. The femur is not
over or under corrected to compensate for the tibial deformity. The tibia
should be corrected separately either during the same or at a later treatment.
Physical therapy(PT) starts within one or two days from surgery. PT should
continue daily throughout the distraction and consolidation phase. It stops
briefly after removal of the external fixator to avoid a fracture through
the regenerate or a pin hole. Once the bone is strong enough it continues.
During the distraction phase one to two formal sessions each day(45-60 minutes
each) with a therapist are required. In addition at least 2 home sessions
per day (30 minutes each) are recommended. The more therapy the better the
potential functional result and the faster the rehabilitation following removal.
Inpatient rehabilitation is often the only practical method of achieving
this quantity of therapy. The philosophy of therapy for lengthening is very
different than for other orthopedic surgical procedures. Following most
orthopedic procedures the patient is at their worst after surgery and gradually
recovers. One week following surgery lengthening patients are at their best.
Thereafter due to the distraction, the muscles become tighter and range of
motion of joints more limited. It is not until the consolidation phase that
the usual pattern of rehabilitation and recovery occurs. One can think of
the lengthening surgery ending at the end of the distraction phase: a surgical
procedure that can be measured in months rather than hours. In the absence
of .a therapy program we will not even consider femoral lengthening.
The majority of the therapy time should be spent obtaining knee flexion and
maintaining knee extension. Passive exercises are the most important during
the distraction phase and passive plus active exercises during the consolidation
phase. Hip abduction and extension are the two important hip exercises.
Strengthening exercises should be focused on the hip abductors and the
quadriceps. Electric muscle stimulation is used on the quadriceps. Upper
extremity strengthening is helpful for use of walking aids and transfers.
Weight bearing is encouraged and allowed as tolerated.
The goal of treatment in this group is to convert the femur into a Type 1.
This requires obtaining union of the pseudarthrosis. Even if failure to obtain
union occurs with the first treatment the pseudarthrosis may be converted
to a stiffer type which can then be treated more easily. To classify this
group into Type 2a it is necessary to do a fluroscopic examination. In the
infant when the ossific nucleus is separated from the rest of the femur by
a large gap of lack of ossification an arthrogram is necessary to demonstrate
differential motion of the head and femur demonstrating a mobile nonunion
and mobile head (fig 12).
In order to obtain union of the head to the femur it is necessary to open
the pseudarthrosis, bone graft it and reorient it. A proximal femur valgus
osteotomy is performed to reorient the nonunion and the coxa vara. The valgus
can be performed either like that of a neck or more like a Schanz pelvic
support osteotomy under the head. If the femur is proximally migrated it
may be necessary to pull down the femur relative to the pelvis as a first
stage. If this is not possible acutely then it should be done gradually.
It is important to extend the fixation to the pelvis in order to neutralize
the forces at the pseudarthrosis site (fig 13). Once the pseudarthrosis is
united the rest of the treatment is as per Type l.
If the hip is determined to be stiff or absent but the knee is mobile,
reconstruction can still be performed combined with lengthening. Motion of
the hip in these patients comes from the mobile pseudarthrosis. The proximal
femoral shaft should not be fixed to the femoral head. If union of the femoral
shaft is carried out to the femoral head the hip will be stiff. To preserve
hip motion but create stability for lengthening and for gait the proximal
femur should be osteotomized into valgus under the femoral head. This type
of hip reconstruction is not performed until age 10-16 yrs.
Prior to this age the femur can be lengthened once or twice. Frequently the
femur is so short that it cannot be lengthened together with hinging of the
knee. In these cases the knee can be locked in extension for the first
lengthening. Either femoral lengthening alone or simultaneous femur and tibial
lengthening may be performed.. Fixation must extend to the pelvis to prevent
proximal migration of the femur (fig 14).
If the femur is large enough then hinging of the knee is performed. This
may wait until the second lengthening. At later childhood or early adolescence
the hip reconstruction osteotomy may be performed in combination with lengthening
If there is a major deficiency of the proximal femoral diaphysis then the
first lengthening is performed to grow the femoral shaft. Obviously the more
deficient the proximal femur the more complicated the lengthening program
becomes. The risk of significantly losing knee motion in this group is very
high. Knee flexion deformity is very common in this group. An arthrogram
should be done to determine if the femoral condyles are round. If the femoral
condyles are round the knee can be extended by soft tissue releases and
distraction. If the femoral condyles are flat the knee should be extended
Obviously as the degree of deficiency increases the appeal of a prosthetic
option of treatment increases. This will be addressed later
Deficiency of the proximal femur with an absent femoral head, greater trochanter
and proximal femoral metaphysis has a mobile pseudarthrosis and a very short
femoral remnant. Despite this some cases have a mobile knee with greater
than 45 of motion and frequently as much as 90 of motion. There is usually
a knee flexion deformity of about 45 present. The treatment option in these
cases should include lengthening or prosthetic reconstruction ( Van Nes or
Symes). The latter is certainly less taxing on the patient and family while
the former requires significant lengthening surgery expertise and multiple
surgeries throughout childhood. To lengthen Type 3a femur deficiency the
femur should be converted to a type 2b. After that treatment is as described
above for Type 2b. The knee flexion deformity in these cases can be addressed
in one of 2 ways. If there is flattening of the articular surface of the
distal femur then the correction of the flexion deformity can be performed
by extension osteotomy of the distal femur. If the condyles appear round
on arthrogram then the knee flexion deformity can be corrected by extension
osteotomy of the distal femur combined with lengthening through the same
osteotomy. Because of the severe discrepancy from this type of deficiency,
a combined femur and tibia lengthening can be carried out. Since the femur
is so short it is impossible to lengthen and allow knee motion at the same
time. The frame is therefore extended across the knee joint without hinges.
The fixator is also extended across the hip to the pelvis to prevent proximal
migration of the femur.
Treatment Type 3a: Diaphyseal Deficiency, knee ROM < 45 degrees
Lack of motion of the knee to < 45 deg is due to malformation of
the lower end of the femur and abnormalities and deficiencies of some of
the soft tissues.. The femoral condyles are flat and the tibia articulates
in flexion with the squared off end of the femur. There is an absence of
the suprapatellar pouch. The quadriceps muscle and tendon are usually present
( they may be absent) although the patella is usually absent. Complete rigidity
of the knee is frequently related to an abnormal ligament running from the
insertion site of the anterior cruciate to the anterior surface of the distal
femoral condyles. Once this ligament is resected the knee can flex. Resection
of this ligament frequently results in 45 deg of knee flexion. Since the
femur is so short it is very difficult to exercise the knee since there is
no proximal lever to stabilize. The best result that can be hoped for in
these limbs with a lengthening program is equalization of limb length, a
mobile hip, a mobile ankle and a stiff knee with at the most 45 deg of motion.
Functionally this is probably not as good as a Van Ness rotationplasty which
offers active motion of the knee. In comparison to a Symes with an above
knee prosthesis the difference is probably the convenience of a passively
mobile knee with the prosthesis compared to the advantages of foot sensation
without the need for a prosthesis following lengthening. In the Van Nes and
the Symes the hip is deficient and unstable.; Ilizarov hip reconstruction
to minimize limp and give better pelvic support can be considered following
Symes and Van Nes. A lengthening program for Type 3b is a very long haul
requiring multiple procedures to achieve limb length equality. It should
only be considered in patients that absolutely refuse one of the prosthetic
options (fig 15), have bilateral disease or phocomelia affecting the upper
The majority of Type 1 CFD require at least 2 lengthenings, As the expected
discrepancy at skeletal maturity increases the number of lengthenings required
to equalize LLD increases. The amount of lengthening that can be performed
in the femur at any one time is dependent on the initial length of the femur.
Generally 4-6cm can be performed safely in toddler (age 2-4) femurs. In children
over age 6 at least 6-8cm is usually possible. In adolescents and young adults
8-12cm may be possible in the femur. Combined femur and tibia lengthenings
allow greater lengthening amounts. Tibial lengthening of up to 5cm can be
combined with the above femoral lengthening amounts. Toddler lengthening
should only be considered in children with a well developed hip joint including
an ossified femoral neck. Toddler lengthening is usually limited to 4-6cm
although we have safely performed up to 8cm in the older toddlers if knee
motion is well maintained. The two main advantages of toddler lengthening
are growth stimulation and reduction of prosthetic needs. Growth stimulation
was seen consistently in our toddler femoral lengthenings. It is a progressive
stimulation in some cases while in others it returns to the previous growth
rate. With a reduction of the leg length difference toddlers are able to
reduce one level of prosthetic limb length equalization. This means going
from a long leg prosthesis to an AFO and shoe lift, AFO and shoe lift to
shoe lift only or shoe lift to no lift. The complication rate in this group
is no higher than in older children.
Discrepancy at maturity:
|1 lengthening over age 6|
2 lengthenings: toddler(<5cm) + age 8-10(<8cm)
1 lengthening: toddler(<5cm) or age 6(<8cm) +epiphyseodesis(<5cm)
|12-16||2 lengthenings: toddler(<5cm) or age 6-8 (6-8cm) + age 10-12 (8-10cm)|
'2 lengthenings: toddler(<5cm) or age 6-8 (<8cm) + age 10-1'2
+tibia (<5cm)during one of the femoral lengthenings
3 lengthenings: toddler(<5cm) + age 8-10 (6-8cm)
+ age 12-16 (10-12cm)
3 lengthenings + epiphyseodesis
The above are some strategy formulations combining the amount of lengthening
possible (safe) at different ages with additional lengthening in the tibia
and/or contralateral epiphyseodesis of the femur. There are more possible
strategic combinations than those listed. The strategy must allow time for
the additional procedures required to correct deformities and instabilities
of the hip and knee.
Lengthening Reconstruction Surgery vs Prosthetic Replacement Surgery
Strategies and methods of lengthening reconstruction surgery(LRS) at our
center is described above. Prosthetic replacement surgery(PRS) refers to
Symes amputation and Van Ness Rotationplasty followed by prosthetic fitting.
The latter two PRS have been used almost indiscriminately for all types of
CFD. The reason for this dates back to the disastrous experience with the
Wagner technique for LRS of these femurs (13). Many patients were worse off
after the reconstruction than if surgery had never been performed (treatment
worse than the disease). In our experience with the above strategies in children
and adults not a single patient thus far falls into that category. We have
not compared our LRS results with PRS results since we do not have a comparable
cohort. Nevertheless our results in 54 patients with congenital short femur
syndrome are: Type la (45 patients) Excellent-32, Good-10, Fair-3 and Poor-0
(result score is based on clinical subjective, clinical objective and
radiographic criteria); Type lb (2 patients) Good- 1, Fair-1; Type 2a (1
patient) Good-1; Type 2b (3 patients) Excellent-l, Good - 1, Fair-1; Type
3a (1 patient Good -1; Type 3b (2 patients) Good-2. Many of these patients
have completed only one lengthening while others have completed as many as
three lengthenings. In a separate study of 70 Ilizarov femoral lengthenings
clinical and radiographic results were compared between congenital,
post-traumatic and developmental cases undergoing lengthening. There was
no significant difference in results based on etiology.
While more authors are recommending LRS, pseudarthrosis and the status of
the hip is used as a primary deciding factor for LRS vs PRS. It should be
emphasized that the hip status does not change after PRS. We argue therefore
that the status of the hip should not be a major deciding factor for PRS
or not. In fact hip procedures used for LRS are useful to stabilize the hip
and improve gait even if PRS is chosen. The status of the knee for us is
the deciding factor to recommend LRS vs PRS. Therefore our absolute indications
for PRS are primarily in Type 3 cases. In Type 2 cases it should also be
considered depending on how good the knee is and how much predicted discrepancy
there is. Type la&b should rarely be considered for PRS, unless there
is a stiff knee associated with these types. Finally in Type 1 CFD, LRS is
so reliable in our hands that PRS should only be considered when psychologic
or socio-economic reasons prevail.
One of the arguments for PRS is the contention that LRS leads to psychologic
scarring and loss of childhood. In our experience LRS if properly conducted
with an appropriate rehabilitation program and surgeries strategically spaced
apart does not lead to obvious psychological scarring to the child. In fact
my experience (having written several letters of recommendation to college
for these children) is that most children develop stronger characters and
a more goal oriented approach to life. It can truly be a growing
experience. LRS is an investment. The child invests part of their childhood
in order to live the majority of their life as an adult with as near normal
an extremity as possible. We try to complete the LRS before the child enters
high school whenever possible so that the formative years of body image at
the time when the children are most self conscious are with both limbs of
equal length and near normal function. In this manner most go through a normal
adolescence. The psychological stress of wearing a prosthesis during adolescence
is not well quantified by psychological profiles performed on these individuals
as adults. Therefore it is difficult to compare LRS vs PRS.
Psych-socio-economic stresses can play a major role in the decision LRS vs
PRS. Single parents, marital difficulties, financial difficulties, drug problems,
behavioral problems, learning disabilities and mental capacity, etceteras
may interfere in the compliance, maturity, and home stability required to
undergo LRS. PRS is easier, more painless and requires far less of the family.
In such situations where the family would find it difficult to comply, or
too stressful for the other family members PRS is the preferable option.
Distance may play a factor too. If the patient is unable to commute to a
center that can provide successful LRS then PRS may be a preferable option.
This also applies to the postoperative rehabilitation required which is an
absolute prerequisite. This problem may be soluble by rehabilitation hospitals
and free care hospitals for children. Finally successful LRS requires a team
dedicated to this type of treatment. It is not a procedure that should be
performed casually or by surgeons inexperienced in the treatment of these
patients. Experience in limb lengthening for other conditions is not sufficient
to know how to successfully lengthen children with CFD. It requires a long
commitment of time on the surgeonís part and on the part of the surgeons
team. It requires appropriate rehabilitation services. If all of these facilities
are not available then LRS should not be considered at that venue. The latter
is perhaps the main limiting factor today in the availability of LRS.
Note from PFFDvsg Webmaster: For more information
The International Center for Limb
Sinai Hospital of Baltimore
2411 West Belvedere Avenue
Baltimore, Maryland 21215 USA