Leg-Length Discrepancy

When a patient is born with legs of unequal length that
are otherwise normal, it is often impossible to know which leg is the
abnormal one. Because the more severe cases clearly involve shortening,
it is appropriate to think of these cases as hemiatrophy rather than
hemihypertrophy. Beals has stated that hemiatrophy and hemihypertrophy
are distinct clinical syndromes, partly because of their associated
anomalies (33). The dysplasia usually involves
the entire limb, with some shortening of all components, and usually is
accompanied by a diminution in girth. Each leg appears to be
genetically programmed to be of a different size (34).

Congenitally short bones frequently show qualitative, as well as quantitative, changes (Fig. 29.12) (35). The congenitally short femur also can show coxa vara, bowing, hypoplasia of the lateral condyle (Fig. 29.13), and external torsion (36). It can be associated with anterior cruciate insufficiency (37,38,39,40), a short or missing fibula (41,42),
and absence of the lateral rays of the foot. Indeed, the congenital
short femur is thought by some to be a variant of proximal focal
femoral deficiency (43).

Congenitally bowed tibias are frequently accompanied by leg-length discrepancy and hypoplastic feet (44,45,46). Askins and Ger (47) found that 24% of patients with congenital constriction bands have leg-length discrepancy; and Garbarino et al. (48) reported short tibias in association with congenital diastasis of the inferior tibio-fibular joint.

A trauma that injures the physeal plate can slow its
rate of growth either by direct injury to the cells responsible for
growth or by formation of a bony bridge that tethers the epiphysis to
the metaphysis. Salter and Harris provided a classification of
fractures of the physeal plate that is useful in anticipating the
effect of fractures on future growth (49). This classification is shown diagrammatically in Figure 29.14.
Fractures can wander through all zones of the plate, but tend to pass
through the zone of cell hypertrophy where the material is weakest and
the amount of material is least. The material in that zone is
cartilage, which is weaker than bone, and because the cells there are
large, the ratio of matrix volume to cell volume is low (Fig. 29.15).
It is important to note that this part of the plate is the site of
conversion of cartilage to bone, but it is not primarily responsible
for growth that occurs by virtue of cell multiplication and matrix
production in the zones nearer the epiphysis.

Because type I and type II fractures do not pass through
the growth zone, they are less likely to interfere with growth than
other fractures are. Both types of fractures, however, may be
associated with a crush injury, which injures the cells by compression.
This mechanism may account for the higher than expected incidence of
growth disturbance in type II fractures of the weight-bearing bones,
such as the distal femur, where growth arrest is found in more than one
third of patients (50). Type III and IV
fractures do, however, cross the growth zone and are therefore more
likely to result in growth arrest. The type IV fracture, in particular,
can result in a bony bridge when the fracture fragment displaces in the
diaphyseal direction (Fig. 29.14). This is one
reason why these fractures must be anatomically reduced. Type V
fractures can occur in isolation or can accompany any of the other
types of fractures. Type V fractures are insidious because they are not
initially recognizable on radiographs and they always demonstrate their
presence by a disturbance of growth, either a shortening or

a
combination of shortening and angulation, usually in the first year
after the occurrence of the fracture. Although the fracture
classifications provide guidelines about the likelihood of growth
arrest, the orthopaedic surgeon must be wary of giving a definite
prognosis for a given epiphyseal fracture until enough time has elapsed
to rule out a type V injury.

The bony bridge that causes a growth disturbance
following physeal fractures usually is discrete and well defined and
lends itself to excision if the fractures are small and peripheral.
Bridge resections usually are limited to those that involve less than
50% of the plate in patients who have at least 2 years of growth
remaining. Even more extensive resections can be considered in very
young children because, if successful, difficult treatment of severe
leg-length discrepancy might be avoided. Resection of a bony bridge
always should be considered if there is significant growth remaining,
even if leg-length discrepancy is already present. The angular
deformity that may also be present because of the bridge can influence
treatment of the discrepancy because both deformities can be corrected
at once.

Osteomyelitis adjacent to the plate can result in the
destruction of physeal cells and disturbance of growth if not treated
early (51). The infection is usually
hematogenous osteomyelitis of the metaphysis, but it can be epiphyseal
in infants and can follow or precede septic arthritis of the joint. The
bony bridge that results from infection is more difficult to treat than
that following trauma because it is not so amenable to resection. The
bridge tends to be larger,

more
central, and less discrete than that following trauma and can even
consist of multiple small bridges. It is difficult to define by
radiograph, is usually more extensive than it appears, and is more
difficult to define during resection. There is the danger that minor
components of the bridge can be left behind because the usual end point
of resection, a continuous line of physis around the resection tunnel,
can be achieved despite incomplete resection of all components of the
bridge.

Infection tends to produce more serious leg-length
discrepancy problems than trauma does because it occurs commonly in
younger children with much growth ahead of them. As with trauma,
angular deformity and leg-length discrepancy can coexist.

Inhibition of growth commonly accompanies weakness or
paralysis of the leg, but the mechanism of this inhibition is not
clear. It may be true that blood flow to the limb is reduced because of
the reduced muscle mass, but this does not necessarily mean that flow
to the plate is also reduced. Venous return results partly from muscle
activity and, therefore, it is conceivable that reduced muscle activity
and decreased pumping effect could reduce blood flow to the limb and
perhaps to the plate. Alternatively, abnormal vasomotor control from
the neurologic abnormality could affect blood flow.

Paralysis and reduced muscle activity may have a more direct effect on the growth rate. The Heuter-Volkmann law (29)
suggests that the growth rate of the physis responds to the compression
forces across it, and this law is frequently used to explain how a
spontaneous reorientation of the physis occurs while contributing to
the remodeling of angular deformities in the immature child. This
mechanism also can explain the decrease in the overall growth rate that
occurs in children with muscle weakness.

The parents of children with cerebral palsy are
frequently concerned about leg-length discrepancy, and minor degrees of
discrepancy can be seen in this condition. More often, however, the
discrepancy is more apparent than real and occurs because of pelvic
obliquity due to hip contractures or asymmetric posturing due to
asymmetric spasticity. It is likely that serious discrepancies do not
occur more often because even dysfunctional spasticity can be
effective, through the Heuter-Volkmann law, in stimulating the physis
to grow.

Leg-length discrepancy can be related to tumors in
several ways. The first way involves destruction of the plate by direct
tumor invasion, the mechanism of invasion in this instance being
similar to that of infection.

The second way involves damage to the plate by irradiation that is used to treat the tumor (52).
Irradiation has a particularly harmful effect because the osteocytes of
the neighboring bone also are killed, and the bone can take many years
to become revascularized and repopulated with healthy osteocytes. The
absence of healthy osteoblasts and precursors can complicate the
treatment of the ensuing leg-length discrepancy by precluding
lengthening procedures through the affected bone. Radiation damage to
regional soft tissues also complicates lengthening procedures (53,54).

The third way that tumors produce leg-length discrepancy
involves the diversion of cartilage cells of the physis into the tumor,
thereby stealing growth potential from the plate. Examples of this are
enchondromatosis, Ollier disease (55), and
osteochondromatosis, all of which are derived from chondrocytes of the
growth plate. Although unicameral cysts usually do not result in
significant leg-length discrepancy, some of the growth disturbances can
result from aggressive attempts to remove the cyst wall when the cyst
is active and immediately adjacent to the plate. Unicameral cysts and
fibrous dysplasia cause leg-length discrepancy because of both growth
inhibition and repeated fractures with minimal displacement that
produce progressive varus.

Because the circulation of the physis is derived from
the epiphyseal circulation, avascular necrosis of the epiphysis
frequently involves the growth plate as well, resulting in leg-length
discrepancy. This may be seen in Legg-Perthes disease and in avascular
necrosis following treatment of developmental dislocation of the hip.
Peterson has reported a case in which a discrepancy resulted from a
temporary but substantial episode of vascular insufficiency during

surgery in an infant (56).
In the case of congenital dislocation of the hip, the effect is
maximized by onset at an early age and by the years of future growth
that are affected but is moderated by the fact that the growth plate of
the proximal femur contributes only approximately 15% of the growth of
the limb. The likelihood of substantial discrepancy has been correlated
with the pattern of ischemic damage to the head and increases with
increasing involvement (57). The fact that the
vascular damage to the epiphysis does not always significantly affect
the physis is indicated by the observation that patients with
Legg-Perthes disease do not usually develop substantial deformity (58). Leg-length discrepancy also has been reported as a complication of catheterization of the umbilical or femoral artery (59), presumably because of impairment of the arterial supply to the physis.

Vascular malformations, particularly when they involve
large portions of the limb, produce growth stimulation that often
involves all growth plates of the limb and not just the ones in
proximity to the tumor or the plates of the involved bone. This
stimulation is seen with hemangiomatosis and Klippel-Trenaunay-Weber
syndrome (68). Stimulation is also seen with
certain nonvascular tumors, such as neurofibromatosis, fibrous
dysplasia, and Wilms tumor, although an increase in circulation could
be the final common pathway to growth stimulation.

Overgrowth of the involved bone is a common feature of
chronic osteomyelitis, presumably because of the increased blood flow
to the limb as part of the inflammation. Infection, therefore, can both
inhibit and stimulate growth. Overgrowth of the affected limb can be
seen in pauciarticular juvenile rheumatoid arthritis (69), particularly in those cases with onset before the age of 3 years (70), and has also been reported in a patient with hemophilia and with chronic knee synovitis (71).

Overgrowth usually is seen following fractures of long
bones in children and also is believed to result from the increased
blood flow to the limb that is part of the healing process (72).
One particularly pernicious example of this effect is the overgrowth of
the tibia and valgus deformity that can follow minimally displaced
proximal metaphyseal fractures (73). The mechanism involves overgrowth of the medial side of the tibial growth plate (74), possibly because of tethering by the fibula or by release of the torn medial periosteum (75,76,77,78).

Overgrowth most commonly occurs following femoral fractures in young children (79,80).
Some studies have reported that the stimulatory effect can last for
years, but it is believed to occur principally during the healing and
remodeling periods in the first 2 years after the fracture (81).
The stimulation has been reported variously to be the greatest in
fractures in the proximal third of the femur, the middle third of the
femur (82), and in those fractures with greater degrees of overriding (83); it can be accompanied by overgrowth of the fractured tibia on the same side. Conversely, Meals (84)
found that the patient’s age and the type and location of the fracture
did not influence the extent of overgrowth, although, inexplicably,
handedness did.

A tape measure is used to measure the real length of
each leg from the anterior superior iliac spine to the tip of the
medial malleolus. The apparent length is measured from the umbilicus to
the tip of the medial malleolus. These two measurements of the
leg-length discrepancy may yield different results, because the
apparent length is affected by pelvic obliquity and hip position. The
leg on the side of the adducted hip appears shorter, but the real
length is minimally affected (Fig. 29.16). The
patient should be undressed completely for this measurement to avoid
tenting of the tape over the clothes. In any case, the tape often tents
over or around the knee, impairing the accuracy of the measurement.
Because different landmarks are used, the lengths of the legs measured
in this way are not the same as those measured by radiograph, but the
discrepancies should correspond closely.

The medial aspect of the joint line of the knee can be
used as a landmark for measuring the segment lengths of the tibia and
femur. Although this method is less accurate than the radiologic
measurement, it is useful for comparison to prevent errors. If the
relative heights of the knees will be a strategy factor, then the
lengths of the individual bones can be determined as well.

It is useful to place blocks under the foot on the short side to lengthen the short leg effectively (Fig. 29.17).
The block height required to produce a level pelvis should correspond
to the measured discrepancy. Blocks also can be used to estimate the
extent of correction that feels best to the patient and provides him or
her with the best correction. This amount can be different from the
radiographic measurement, and it may indicate that the goal of
treatment should not be exact correction of leg length. This assessment
is most useful in the mature patient whose discrepancy will not be
changing with growth. Although it only measures the present length and
does not indicate the desired amount of correction, it can be helpful
in immature patients to indicate that exact correction is not the best
goal. Carrying this idea further, the patient can be provided with a
temporary shoe lift and can be assessed after a period of ambulation.
These techniques are especially useful for patients with complex
deformities, because they take into account the combined effects of
asymmetric feet, angular deformities, contractures, pelvic obliquity,
and spinal balance.

Several factors other than leg length must be assessed
because they affect the measurement of leg length or influence the
final outcome of the patient’s treatment.

The examiner must remain aware that knee and hip flexion
contractures tend to shorten the leg, an equinus contracture of the
ankle tends to lengthen it, and pelvic obliquity also tends to affect
apparent length. The term pelvic obliquity has a broader meaning here
than when used by spinal surgeons, who use it to refer to the relation
of the position of the pelvis to that of the spine (85).
In the context of leg-length discrepancy, it refers to the relation of
the position of the pelvis to that of the legs, and it is affected both
by adduction and abduction contractures of the hips and by spinal
deformity. An adduction contracture of the hip causes the leg on the
affected side to appear short and to be functionally short, whereas an
abduction contracture has the opposite effect (86).
This situation is common in patients with cerebral palsy and those with
the residuals of poliomyelitis. A difference between the measured real
and apparent discrepancies indicates that pelvic obliquity is present.

Hip contractures are particularly important because
adduction contractures produce a functionally short leg and abduction
contractures produce a functionally long leg (87).
These effects occur even in the absence of an actual discrepancy as
measured by a scanogram. Conversely, in a patient with true shortening,
a limitation of ipsilateral hip abduction or other side hip abduction
will prevent him or her from compensating easily and will exaggerate
the effect of the discrepancy.

Angular deformity must be assessed because it affects
the measurements of leg length and influences the final outcome if it
is to be corrected later. Joint stability must be assessed because it
pertains particularly to the risks of leg lengthening. Femoral
lengthening is contraindicated in the presence of hip instability.
Congenitally short femurs always are associated with laxity of the
anterior cruciate ligament (37) and with hypoplasia of the lateral femoral condyle (Fig. 29.13)
that predisposes to posterolateral subluxation of the tibial plateau.
Lengthening of the tibia can be contraindicated in the presence of an
unstable ankle or a useless foot which might be better handled by
amputation and prosthetic fitting.

Spinal deformity and balance should be assessed. If
there is stiff suprapelvic obliquity such that the axis of the trunk
cannot be made perpendicular to the transverse axis of the pelvis, an
equalization of leg lengths will cause an imbalance of the trunk, and
some modification of that goal will be necessary. Adduction and
abduction contractures of the hips produce infrapelvic obliquity, with
apparent and functional leg-length discrepancy (88).

Weakness and the need for bracing must be assessed
because leg-length discrepancy in patients with paralysis or weakness
is usually best handled by undercorrection, leaving the weak leg short
to facilitate swing-through, particularly if the leg is braced with the
knee locked in extension. Patients who require bracing of the short leg
to walk can have their leg-length discrepancy corrected in the brace
and may not require surgical correction at all.

Finally, the concerns, compliance, and emotional state
of the parents and patient must be taken into account. This aspect is
particularly important when lengthening is being considered, because
this is a long and difficult process requiring understanding and
cooperation by all involved. Despite excellent education and
preparation, parents and patients always underestimate the challenge of
dealing with the duration of lengthening and the later restriction of
activities. If there is a lack of understanding or a suggestion of poor
compliance, then another approach may be more appropriate. The surgeon
constantly must be aware that patients frequently express concerns
about function when they are actually concerned about cosmetic effect,
which may be less important when compared with the risks of surgery.

Several methods exist for the radiologic measurement of
leg length. These methods are more accurate than clinical methods, but
each has its advantages and disadvantages. The nonexistent, ideal
method would allow the hip and ankle to be viewed, minimize radiation,
use only one exposure, use a film of convenient size, demonstrate
angular deformity, have no magnification, give true readings from a
scale on the film, and be inexpensive.

The bony landmarks used are the top of the femoral head, the medial femoral condyle, and the ankle. The ankle

mortise is slightly saddle shaped, and the midpoint of the saddle can
be easily identified. Although these techniques allow the measurement
of the femur and tibia individually, these values are not required in
analyzing and predicting growth, and the length of the entire leg
suffices.

The surgeon cannot rely on measurements recorded in the
patient’s record but must review all radiographs before performing
surgery to check their accuracy and reliability. There is the
possibility that radiographs taken over the years have been read by
different individuals using different techniques and landmarks. It is
also possible that the scales have been misread or that arithmetic
errors have been made in determining lengths and discrepancies.

There are four radiologic techniques that measure leg
lengths directly and others that provide useful information. The
terminology is confusing because names used for these techniques are
inconsistent in the literature and in use. The term “scanogram,” for
example, was derived from the technique of split scanography, in which
the x-ray beam was tightly collimated to a thin transverse slit that
exposed the film as the x-ray tube was moved from one end of the leg to
the other. The principles are more important than the terminology.

The teleoradiograph (Fig. 29.18)
is a single exposure of both legs on a long, 35 cm × 90 cm (14 in × 36
in) film. It is taken from a 2-m (6-ft) distance, usually with the
patient standing and with a radiopaque ruler placed on the cassette. It
has the advantage of showing angular deformities and of using a single
exposure, but it produces a radiographic film that is inconvenient to
handle and measurements that are subject to magnification because of
parallax of the x-ray beam. There is no significant difference between
measurements of the leg taken supine and those taken standing (89).

The orthoradiograph (Fig. 29.19)
avoids the magnification factor by taking separate exposures of the
hip, knee, and ankle so that the central x-ray beam passes through the
joints, giving true readings from the scale (90).
However, the orthoradiographic film is still cumbersome, and the need
for multiple exposures introduces the risk of error if the patient
moves.

The scanogram (Figs. 29.20 and 29.21)
avoids magnification in the same way but reduces the size of the
resulting film by moving the film cassette beneath the patient between
exposures (91). This technique is preferred in
children older than 5 or 6 years who can be compliant with instructions
not to move, because it gives true measurements without

magnification, but younger children are better assessed using the teleoradiograph.

Positioning for the scanogram must be modified for
patients with contractures of the hip or knee. Patients with hip
flexion contractures can have accurate measurements taken in the
reclining position. In those with only knee contractures, the femur can
be measured in either the lateral or the prone position, and the tibia
can be measured in the lateral position. Assessment of both bones can
be made on one x-ray film in the lateral position if two rulers are
used, one parallel to each bone. If a hip contracture is also present,
the femur must be assessed in the lateral position. The scanogram of
both femur and tibia can be done in the lateral position.

Digital radiography can be used to measure leg length (92,93).
Computed tomography scan can be used to measure the distances between
points on the radiographic film, reducing errors from angular deformity
(94). If the examination is done specifically for this purpose, the cost is comparable with more traditional techniques (95,96,97), multiple sections are unnecessary, and the radiation exposure is less, especially with microdose techniques (98).

Whatever technique is used, it is important to be
consistent and to not mix true and magnified measurements when
analyzing data to determine the timing of surgery. Because errors are
possible with all of these techniques, the resulting measurements
should be compared with, and should correspond with, the clinical
measurements.

An anteroposterior film of the pelvis and hips, taken
with the patient standing and the legs straight (on blocks if
necessary), is occasionally useful in assessing the combined effect of
leg-length discrepancy, angular deformity, and asymmetry of the foot
and pelvis because it more closely approximates the functional
discrepancy. The leg-length discrepancy is calculated from the heights
of the femoral heads from the floor, by taking into account the height
of the blocks. This assessment can supplement the clinical examination
done with blocks.

It is important for the surgeon to integrate all the
information from the various methods of measuring leg length while
preparing a treatment strategy.

All methods of estimating skeletal age involve comparing
radiographs of the patient with standards in an atlas. Methods have
been described using the bones of the pelvis and hip, the knee, and the
hand and wrist (99,100).
The estimation of skeletal age is only moderately accurate and is the
weak link in the techniques of analysis and prediction of growth.

The Greulich and Pyle method is used universally in this context. Their atlas (20)
consists of reproductions of radiographs of the left hand and wrists of
boys and girls that the authors considered typical for the stated
skeletal age. To estimate the skeletal age of the patient, a radiograph
of the left hand and wrist is taken according to the technique
described in the atlas, and this film is then compared with the
standard radiographs of the appropriate gender according to qualitative
(e.g., the visibility of the hook of the hamate) and quantitative
(e.g., the degree of conformity of an epiphysis to its metaphysis)
criteria. The standard that most closely matches the patient’s
radiograph is taken as the skeletal age.

This technique has certain deficiencies. The first is
that in some parts of the atlas the standards represent skeletal ages
that are far apart, with a gap as great as 14 months. There is
therefore a large standard error built into the technique. With
practice, it is possible to interpolate between standards, but this
practice is difficult, and studies have shown significant interobserver
and intraobserver errors (101). A second
problem is that some children do not follow the same orderly succession
of maturity indicators shown in the atlas, and an arbitrary choice must
be made in assigning a skeletal age. Third, some children with
leg-length discrepancy of congenital origin also have congenital
anomalies of the hand and wrist, making it impossible to reliably
compare their radiographs with the standards. Finally, one of the
features of almost all atlases of skeletal ages is that the mean
skeletal age of a sample of similarly aged children is equal to their
chronologic age so that the skeletal age is, in fact, the best possible
predictor of chronologic age. That is, of a random sampling of boys or
girls of the stated chronologic age, half would be more developed and
half would be less developed than the standard radiograph. Greulich and
Pyle, however, in selecting standards

for
their atlas, did not follow this principle exactly, and in some cases
they selected radiographs that they believed were more representative.

The Tanner-Whitehouse method (102)
is similar in that it uses radiographs of the hand and wrist but was
developed using modern computerized mathematical procedures. It adds a
level of refinement and accuracy to the Greulich-Pyle technique by
defining and showing examples of successive stages of development of 20
specific bony landmarks in the hand and wrist (Fig. 29.22).
The same standards are used for boys and girls. The patient’s
radiograph is compared with the standards, and a letter score is
assigned to each of the 20 landmarks. The letter score then is
converted to a numeric score by consulting a table for the appropriate
gender. The sum of these 20 scores represents the level of skeletal
maturity attained by the patient, and it can be converted to years and
months with a much smaller standard error than with the Greulich-Pyle
technique. Of special interest in dealing with leg-length discrepancy
is that the bony landmarks can be divided into two groups. Tanner and
Whitehouse provide tables for including either the 12 long bone
standards of the hand or the eight cuboid bone standards of the wrist
in the assessment. When these two approaches give different results, it
is reasonable to assume that the long bone standards give a skeletal
age that is more pertinent to the growth of the long bones of the leg.
The concept that the long bones are more important than the cuboid
bones in the context of leg-length discrepancy can be useful, even with
the Greulich-Pyle atlas, in allowing the resolution of difficulty in
selecting the appropriate standard for a given radiograph.

This method is more cumbersome and time consuming than
the Greulich-Pyle method. Its use in the context of leg-length
discrepancy is problematic because this method and the Greulich-Pyle
method can produce different skeletal ages from the same x-ray. This is
surprising and probably is related in part to the fact that different
populations and different techniques were used to develop the standards.

The relation between skeletal age and leg length is not
as reliable as we would like it to be, and we must accept the relative
inaccuracy of skeletal age estimation. It is still possible to make
acceptably accurate predictions.

Shortening through the proximal femur at the level of
the lesser trochanter with blade plate fixation has the advantage of
being proximal to most of the quadriceps origin and therefore not as
disadvantageous to the knee as shortening in the midshaft is. Patients
recover strength and the ability to climb stairs more quickly. This
approach leaves a large scar on the lateral thigh and requires a second
later operation of moderate magnitude to remove the plate.

Winquist (156) and Winquist and Hansen (157)
pioneered a technique that involves the use of a special set of
instruments designed to allow the procedure to be performed entirely
from within the medullary cavity and without any direct approach to the
shaft of the femur. The bone is cut by a special eccentric cam saw that
is passed down the shaft and cuts through the cortex from within. The
size of the saw required is determined by the outside diameter of the
bone, and the femoral shaft is first reamed to an internal diameter
that is sufficiently large to accept it. The distal cut is made first,
and a second cut is then made more proximally at the precise location
to give the desired amount of correction. The cuts should be placed so
that the cylindrical fragment is removed from the isthmus of the femur
where the internal diameter is least, because this site provides the
best fixation to the rod both proximally and distally. The cylindrical
piece of bone can be cut into two sections using a special hook-shaped
reverse-cutting osteotome, and the pieces are pushed aside. The gap can
then be closed over an intramedullary rod to provide rigid internal
fixation. Locking at both ends maintains shortening and rotation. A
second operation is later required to remove the rod.

The technical complications of this procedure usually
result from less than rigid fixation of the fragments, usually due to
inadequate reaming without locking. It is apparent, in using an
unlocked nail, that if a 5-cm segment of bone is to be removed and 2.5
cm of fixation is desired both proximally and distally, then the bone
has to be reamed until 10 cm of the shaft is reamed to the minimum
diameter. This in itself can be a problem because the cortex can become
very thin at the level of the osteotomies, especially because the
reaming is usually eccentric, and can lead to fracture of the shaft.
Less than rigid fixation can lead to loss of rotational control and
opening of the shortening gap, two problems that are difficult to
control without locking. Less reaming is necessary for a locked nail,
but in this case, the reduced internal diameter of the canal may not
allow passage of a cam saw large enough to cut completely through the
cortex, and a percutaneous osteotomy may be required to complete the
cut. Nevertheless, the benefits of control of position with the locked
nail far outweigh this disadvantage, and locking is desirable in such
situations.

Acute respiratory distress syndrome has been reported
during or following closed intramedullary shortening and has been
believed to be the result of fat embolization caused by reaming (158).
Possible preventative measures include venting of the distal metaphysis
and use of reamers with flutes sufficiently deep to allow the unimpeded
egress of reamed material, but the effectiveness of these measures has
not been demonstrated. It is not advisable to use this technique until
this serious complication can be avoided with certainty.

Although this is an appealing technique, it requires
familiarity with the instruments, is technically demanding, and has
best results in experienced hands. The major disadvantages are the
technical complications and risk of respiratory distress syndrome, as
noted in the preceding text, and the significant quadriceps weakness
that results. Patients with greater shortening require 6 to 12 months
to regain normal knee control and function (159).
It leaves a small, cosmetically acceptable scar, and the later
procedure to remove the rod is of lesser magnitude than that required
to remove a blade plate.

The best solution for a child discovered early to have
growth inhibition would be to stimulate his growth to achieve a normal
level. Many techniques have been used in attempts to accomplish this,
but none has been successful enough to be clinically useful. On the
basis of the concept that the periosteum acts as a tether inhibiting
growth, circumferential release of the periosteum has been assessed
both clinically and experimentally and has been found to stimulate
growth (160,161). A variety of foreign materials have been implanted next to the growth plate (162), but the stimulation, if it occurred, was too little and too short-lived to be of use. Sympathectomy (163,164) and surgically constructed arteriovenous fistulae (165,166)
temporarily stimulate growth, presumably by altering the circulation to
the physis, as does stripping and lifting the periosteum by packing
bone beneath it (60,61,167).
However, these techniques have little or no clinical significance.
Electrical stimulation inconsistently stimulates physeal growth, but
even the maximum effect is insufficient to correct clinical
discrepancies (65).

As desirable as this approach might appear, there is no
method of growth stimulation available at this time that is useful in
the treatment of leg-length discrepancy.

Lengthening an abnormally short leg would, at first glance, appear to be the preferable method of dealing with leg-length

discrepancy because it is a corrective procedure. It involves operating
on the abnormal leg to correct its abnormality, as opposed to
epiphysiodesis and shortening, which make the normal leg abnormal and
only compensate for the discrepancy.

Lengthening was first mentioned in the 18th century, in
an account of injuries sustained in battle by Ignatius of Loyola. It
was next reported by Codivilla at the beginning of the 20th century (168).
A number of advances have been made in the surgical technique, or the
lengthening apparatus, or both, and each change has been greeted with
hopes that it would solve the problems associated with lengthening (169).

Lengthening is a procedure of last resort and is
reserved for those situations in which other methods of correction are
inappropriate. Lengthening is usually not appropriate for patients
requiring correction of less than 6 cm because in these cases other
procedures of lesser risk and morbidity can be used. A reasonable goal
of lengthening for most patients is less than 10 cm for the femur and 7
cm for the tibia. Patients requiring large corrections may require
simultaneous lengthenings of femur and tibia, repeated staged
lengthenings of the same bone (170), or
supplementary shortening procedures on the long leg. There is a
threshold, at approximately 15 to 20 cm, where the risks outweigh the
benefits, and lengthening is abandoned in favor of amputation and
prosthetic fitting.

Since Codivilla’s report, there have been many techniques described for lengthening the leg. These have included step cuts (171), periosteal sleeves (172), onlay cortical grafts (173), slotted plates (174), intramedullary rods (175), and other internal and external devices for gradual controlled lengthening (176,177,178,179). Transiliac lengthening has been performed (180)
and may be indicated in cases with both infrapelvic asymmetry and
decompensated scoliosis cases requiring concurrent hip stabilization (181,182). Techniques of instantaneous lengthening of the femur (116,155,183,184) and tibia (185)
have been reported but have not gained widespread support because the
amount of length to be gained is limited. Simultaneous shortening of
one femur and lengthening of the other with the excised bone segment
from the other-side leg has been recommended (186,187).
The Anderson device, using large pins and an external fixator with
threaded rods for lengthening, became widely used but confined the
patient to bed (Fig. 29.30). Some of the older
methods persist in nonindustrialized nations (e.g., double oblique
osteotomy followed by elongation by balanced skeletal traction) (188).
Many of these methods became obsolete with the introduction of the
Wagner device and, later, the Ilizarov and Orthofix devices.

Historically, new lengthening technology has been
adopted enthusiastically by the orthopaedic community and has been used
extensively with great optimism. However, the complication rate has
remained high, and the patient’s course difficult, leading to the
realization that the human leg has not made similar advances and that
lengthening is still a difficult matter for both the surgeon and the
patient.

For two weeks following the application of the
lengthening device the patient will require intensive education and
observation. This allows time for the initial delay and for the patient
and parents to understand and become comfortable with the lengthening
mechanism, pin site care, an exercise program to maintain mobility and
to attain ambulation with weight bearing. Inpatients undergoing
lengthening should be examined daily for blood pressure, neurologic
status, and range of motion. The reading from the scale of the
lengthener should be recorded daily and should be checked to be sure
that the lengthening process is going according to plan. These
assessments also should be made at weekly or biweekly visits after
discharge. Radiographs are taken at intervals of 2 to 4 weeks to
evaluate alignment and the quality of bone in the lengthening gap
(i.e., the regenerate). Ultrasonography can be used instead of
radiographs to measure the lengthening gap (189,190). The rate of distraction can be modified according to clinical progress or radiologic appearance.

Maintaining motion is extremely important during the
lengthening procedure. Patients and parents should be instructed in a
home exercise program, and the patient’s range of motion should be
monitored regularly. Stopping the lengthening should be considered if
limitation of motion that is resistant to a more intensive motion
program develops. Wagner recommended that distraction should not be
performed on any given day if the patient cannot achieve 60 degrees of
flexion (191). My personal guidelines are to
discontinue lengthening of the femur when a knee flexion contracture is
greater than 10 degrees or knee flexion is less than 30 degrees.
Lengthening can be started again if those guidelines are later met
before consolidation of the regenerate prevents it. All patients regain
flexion in the first year after lengthening (192),
and it appears that maintaining extension is more important. Patients
are allowed to ambulate with full weight bearing, with aids if
necessary. The pin sites are cleaned twice daily and, if necessary, the
stab wounds are elongated weekly under local anesthesia to prevent
tenting and ischemia of the skin, which could lead to pin tract
infection. Cutting pins and small wires circumvent the need to release
pin sites.

Distraction is discontinued either when the goal has
been achieved or when an unresolvable complication, usually loss of
motion, supervenes. The device is retained until radiographs show
consolidation and suggest adequate strength of the regenerate bone. In
the consolidation period, dynamization of the device is considered to
be important to subject the bone to cyclic longitudinal loading to
stimulate bone formation. If the bone in the lengthening gap is slow to
consolidate, the device can be shortened to put the bone under
longitudinal compression, either leaving it somewhat shortened or
lengthening it once again when the regenerate responds. Valid objective
guidelines for what constitutes adequate consolidation for removal of
the lengthening device have not been established. Findings such as
corticalization with three cortices visible on two radiographs and the
appearance of a medullary cavity are considered to be signs of adequate
strength, but the decision to remove the device is still empiric.

It is possible to protect the tibia externally with a
cast or brace after device removal, allowing removal from the tibia
earlier than from the femur. In addition, the mechanical and anatomic
axes of the tibia are collinear, and the bone is subject mainly to
compressive forces. This is not the case for the femur, in which the
regenerate bone, especially for proximal lengthenings, is eccentric and
is subject to bending loads. Patients should be restricted from violent
body-contact sports for a long time, and perhaps until the bone is
radiologically normal, because fractures through the lengthening gap
have been reported years later.

Wagner developed a lengthening device that was first used in North America in 1973 (Fig. 29.31) and appeared to offer several advantages over older devices, such as the Anderson device (Fig. 29.30) (54,191,193,194,195).
It is unilateral and uses half-pins instead of through-and-through
pins, thereby facilitating application to the femur. It is small and
light, and it was the first device to allow the patient to be
ambulatory, whereas older devices required confinement to bed. The
patient can perform the lengthening procedure simply by turning a knob
at one end. It is adjustable in varus-valgus and anteroposterior
angulation but not in rotation (Fig. 29.31), so that minor angulation

that occurs during lengthening can be corrected without removing the pins.

The Wagner technique involves at least three surgical
procedures. The first involves performing an osteotomy, releasing soft
tissues if necessary, and applying the device. At the end of the
lengthening phase, a second procedure involves bone grafting the
lengthening gap, plating the bone, and removal of the lengthener.
Months or years later, when the bone has achieved sufficient strength,
a third procedure is done to remove the plate. There is a prolonged
period of restriction of activities to protect the bone and to avoid
late fracture.

Although the Wagner device retains its simplicity and
utility, it has become clear that it is ill-advised to plate the bone
in the presence of contaminated pin tracts. Therefore, the technique
has largely been replaced by newer methods that involve neither plating
nor grafting and appear to reduce the complication rate. The device,
however, remains simple and satisfactory and can be used with the newer
biologic principles.

De Bastiani et al. have developed a lengthening device, the Orthofix device (Fig. 29.32), which is applied to the bone with two sets of conical screws (196,197). It has evolved into a lengthener in which pin blocks move along a rail placed beside the pins (Fig. 29.32).
This arrangement allows the pin sets to start close together and
provides long excursion. It allows the use of up to three screws in
each set, which is an advantage, especially in the proximal femur. It
is technically similar in operation to the Wagner device but, although
it offers more stable fixation to bone, it has a more cumbersome method
of elongation and is not easily adjustable once in place. Encouraging
results have been reported with this method, with a complication rate
similar to that of other techniques (198).

Distraction epiphysiolysis was pioneered by Ring and more recently reassessed by Monticelli, Spinelli, and others (199,200,201,202,203,204,205).
It is achieved by applying a distraction force across the physis until
it fractures. Lengthening can then be obtained by gradual distraction.
This method has the disadvantages that the lysis is sudden, painful,
and not well tolerated and that the physis can be injured, thereby
compounding the leg length inequality (206,207). The complication rate is high (208)
and, if it is to be used at all, it should be reserved for children who
are very near the end of growth to minimize the consequences of physeal
damage.

Ring fixators have been developed (203,205,209,210,211,212) (Figs. 29.33 and 29.34). They are more complex than the

Wagner and Orthofix devices but are also more versatile in that they
lend themselves to the correction of complex deformities. They can
control more than two segments (213),
can extend across joints, and can be used to translate segments of bone
in the treatment of congenital pseudarthrosis and acquired absences (209).
Fixation is accomplished by tensioned through-and-through wires
attached to complete or partial rings. Unwillingness to use
through-and-through wires in the proximal femur has led to the
development of half-pins, which are now gaining favor at all levels.

In the traditional application of any of the external
lengthening devices, the device is responsible for both maintaining
alignment and achieving distraction. Numerous unsightly scars result
because of the multiple percutaneous pins or wires used to achieve
sufficient stability and because they must be left in place for a
prolonged period until the bone is strong. In 1956, Bost and Larsen
introduced the concept of lengthening over an intramedullary rod (175).
The rod serves to maintain alignment during both the distraction and
the consolidation phases, and the external device serves only to
achieve length (214). In this way, the number
of percutaneous tracts can be reduced, the external device can be
removed at the conclusion of lengthening, and the complication rate,
particularly the rate of regenerate fracture, may be reduced (215). Lin described two cases of 15 cases of lengthening over a rod that required bone grafting (216), so the effect of rodding on osteogenesis remains unclear.

Lengthening over the rod has the disadvantages, compared
with other lengthening techniques, that alignment of the anatomic axis
of the femur cannot be changed and that the mechanical axis of the leg
cannot be controlled. Because the anatomic axes of the femur and tibia
are not collinear, elongation in this manner increases valgus of the
mechanical axis of the leg. Whether or not this is a problem depends on
the initial configuration of the leg. The correction, if necessary, of
preexisting angular deformity or

the valgus produced by the lengthening must be performed at another time, when the rod is no longer in place.

There has been reasonable hesitation in using this
approach because of the fear of producing a serious intra-medullary
infection with the intramedullary foreign material in continuity with
the exterior through the pin tracts. Early experience with this
technique is varied. Some studies report a low infection rate and
recommend the technique (217,218), but others are less enthusiastic, mainly because of a significant rate of infection (219,220,221).
Paley studied a series of 29 cases with case-matched controls and
reported a reduction of time in the external lengthener by 50% and a
2.2 times faster recovery of knee motion. There were no infections and
no fractures, compared to six regenerate fractures in the control group
(218).

Wu et al. performed acute lengthenings over a locking rod but achieved a mean lengthening of only 2.8 cm (222).

The application of this technique is limited to bones
that are straight enough to accept the rod and to children who are old
enough not to be at risk for avascular necrosis of the femoral head.

Because most of the serious complications of lengthening
have to do with infection resulting from pins that traverse the skin
and fractures through the regenerate, there has been much interest in
the development of an elongating intramedullary rod that would avoid
both of these types of complication. Lengthening by motorization (223) and by electronically induced shape memory actuation (224)
have been used experimentally, but reports of mechanical ratcheting
devices are starting to appear in the clinical literature (225,226,227,228).
Rotational movement of the distal femoral segment in relation to the
proximal segment operates an internal ratcheting mechanism and, in the
case of the Albizzia nail, 15 such movements achieve 1 mm of
lengthening. Significant complications have been reported, especially
in the form of mechanical problems with the lengthening (226,227), but the concept is promising.

Several groups have been working in parallel over the
last 3 decades to develop not only improved devices but also improved
concepts and methods of lengthening. In particular, Ilizarov and De
Bastiani have contributed new concepts and a new understanding of the
biology of lengthening that are more important contributions than the
devices themselves (196,197,209,210,211).
Because the biologic principles are not device-specific, they are
considered here in principle. Then, with that foundation, technologic
issues are considered.

Simple lengthenings can be accomplished by a number of
devices, none of which has a clear advantage over another, except that
it might be argued that the simplest device capable of solving the
problem is preferable. Gradual correction of complex deformities can
present demands that can only be met by a more versatile device such as
the Ilizarov device.

All studies of leg lengthening, regardless of technique, have reported high complication rates (165,236,240,241,242,243,244,245,246,247,248,249,250,251).

Most studies report more complications than there are patients, and
many patients do not reach their anticipated lengthening goals with
uncompromised function. The complication rate, however, is related to
the amount of lengthening (245), and if the goal is modest the rate is reduced and the proportion of patients reaching their preoperative goals is increased (117).
Ghoneem et al. found that all of the studied children who had undergone
lengthening had a normal psychological score; 92% had no limitations in
daily activities, and 81% were satisfied with the overall result (118).

Deformity due to soft tissue tension can occur during
lengthening, and angular deformity of the tibia has been related to the
degree of elongation (252). The Ilizarov
technique has been proposed to supplant amputation and prosthetic
fitting in fibular hemimelia, but Cheng et al. report that
uncontrollable deformity of the tibia and ankle joint was encountered (253).
Ultimate alignment may be improved by using arthrography to delineate
the joint surfaces when applying the lengthening device (254).

Hypertension can be seen during lengthening and can occur suddenly (255,256,257).
The mechanism is not clear, but it resolves dependably with shortening
of the lengthening gap and may not recur when lengthening is resumed.

Pin tract inflammation is common and is of little
consequence in itself. True pin tract infection usually responds well
to local care and systemic antibiotics. It can, however, be elevated to
a serious complication if the pin tract infection contaminates a
surgical procedure.

Mechanical failure can occur in several ways. Pins can
break or loosen and require replacement. Fractures through the
lengthening gap (or deformation of the bone in the lengthening gap) can
occur early or late in the procedure. It can occur soon after removal
of the device, indicating that it was removed too soon, or can occur
late, while the bone is still remodeling and has not yet regained its
normal strength (258). In my series of Wagner
lengthenings, there was one fracture 8 years after the lengthening,
which suggests that the bone is extremely slow to remodel and regain
full strength and is probably weak even after it looks normal
radiologically. Long-term follow-up is important in assessing the
results of leg lengthening because complications can occur late.

Nakamura et al. have examined the diameter of the bone
in the lengthening gap and concluded that axial loading while the
device is in place may not be enough to increase the diameter but that
the diameter increases dependably after removal of the device (259).
This finding has profound implications for the strength of the bone
because strength in bending varies as the fourth power of the diameter.
Indeed, there is a report substantiating this conclusion (260).

Subluxation of the knee can occur during femoral lengthening, especially in patients with congenital shortening (261,262).
The subluxation appears to be a posterior subluxation of the lateral
tibial plateau and is always preceded by a loss of extension of the
knee. Dysplasia or absence of the anterior cruciate ligament and
hypoplasia of the lateral femoral condyle are usually found in
association with congenitally short femurs (263,264)
and contribute to this complication. Routine lengthening of the lateral
structures (the biceps tendon and iliotibial band), determined
maintenance of knee extension, and avoidance of continued distraction
in the face of a knee flexion contracture will prevent knee subluxation.

Dislocation of the hip has been reported; it can occur
even in the early stages of distraction and tends to be associated with
previous hip surgery and residual instability (265).

Delayed union is difficult to define in the context of
lengthening, but there is no doubt that certain cases take
significantly longer to consolidate than others. True nonunion occurs
and is similar in appearance and treatment to posttraumatic nonunion,
except that it can occur in narrowed and spindle-shaped regenerate so
that it is fragile even when united.

It appears that the complication rate is affected by the
quality of the underlying bone, but inconsistently so. For example
Naudie et al. showed that the complication rate is higher in patients
with an underlying bone disorder (266), but
patients with achondroplasia seem to have a lower complication rate.
Reasonable lengthening expectations in femoral hypoplasia are unclear (267).

Nerve or artery damage during lengthening can result
from stretching, entrapment by tense tissues, movement of wires or pins
through the tissues, or direct trauma (268). Nerve conduction is affected in one third or more of patients undergoing lengthening (269,270).

Karger et al. found that femoral lengthenings were more
prone to complications than were tibias and that the complication rate
increased with lengthenings exceeding 25% of the initial bone length (271).
We found that the complication rate is significantly higher in children
younger than 8 years than in older children, and this may reflect the
inability of younger children to understand, or remain motivated to
comply with, the instructions therapists and surgeons (262).
Conversely, Noonan et al. found that the complication rate in both
femoral and tibial lengthening increased in patients older than 14
years (236). To some extent, the amount of
possible lengthening depends on the length of the bone to start with,
and this is another reason to perform lengthening only in older
children whose bones are nearing mature length.

Lengthening affects the subsequent growth of the limb.
Price and Carantzas have reported a case of severe growth retardation
in the lengthened limb (272). Viehweger has
found that lengthening the tibia by more than 14% slows down its growth
but that this has a negligible effect on the ultimate clinical result
because it is compensated for by femoral overgrowth (273).

Because of the frequency of concerns, questions,
problems, and complications that surround the care of patients
undergoing lengthening, it facilitates their management to form a
lengthening team in a program with shared responsibilities. A nurse,
physical therapist, social worker, and a skilled technician are
important members of the

team
and should join the surgeon in preparing patients and families for the
lengthening procedure. The team can respond to ongoing needs and offer
support during the lengthening procedure. Families and patients never
fully appreciate the depth and breadth of the hardship they will face,
and they will require more support than is needed for most orthopaedic
cases.

There is reason to believe that the complication rate from lengthening is improving with modern techniques and understanding (Figs. 29.36 and 29.37) (274,275).