Gait and Posture 15 (2002) 195– 206

www.elsevier.com/locate/gaitpost

Review

Leg length discrepancy

Burke Gurney *
Di6ision of Physical Therapy, School of Medicine, Uni6ersity of New Mexico, Health Sciences and Ser6ices, Boule6ard 204, Albuquerque, NM
87131 -5661, USA
Received 22 August 2000; received in revised form 1 February 2001; accepted 16 April 2001

Abstract

The role of leg length discrepancy (LLD) both as a biomechanical impediment and a predisposing factor for associated
musculoskeletal disorders has been a source of controversy for some time. LLD has been implicated in affecting gait and running
mechanics and economy, standing posture, postural sway, as well as increased incidence of scoliosis, low back pain, osteoarthritis
of the hip and spine, aseptic loosening of hip prosthesis, and lower extremity stress fractures. Authors disagree on the extent (if
any) to which LLD causes these problems, and what magnitude of LLD is necessary to generate these problems. This paper
represents an overview of the classification and etiology of LLD, the controversy of several measurement and treatment protocols,
and a consolidation of research addressing the role of LLD on standing posture, standing balance, gait, running, and various
pathological conditions. Finally, this paper will attempt to generalize findings regarding indications of treatment for specific
populations. © 2002 Elsevier Science B.V. All rights reserved.

Keywords: Leg length discrepancy; Low back pain; Osteoarthritis

1. Introduction LLD (FLLD) defined as those that are a result of

altered mechanics of the lower extremities [12]. In addi-
Limb length discrepancy, or anisomelia, is defined as tion, persons with LLD can be classified into two
a condition in which paired limbs are noticeably un- categories, those who have had LLD since childhood,
equal. When the discrepancy is in the lower extremities, and those who developed LLD later in life. In terms of
it is known as leg length discrepancy (LLD). LLD is a functional outcomes such as gait, persons who have
relatively common problem found in as many as 40 [1] developed a LLD later in life are more debilitated by
to 70% [2] of the population. In a retrospective study, it LLD of the same magnitude when compared to persons
was found that LLD of greater than 20 mm affects at who have had LLD since childhood [13].
least one in every 1000 people [3]. The effects of LLD There is disagreement regarding the role LLD plays
on function and the magnitude of LLD warranting in musculoskeletal disorders and the acceptable amount
treatment have been subjects of controversy for some of LLD necessary to warrant treatment. Some investi-
time. gators have tried to quantify a significant LLD, accept-
Studies have investigated the effects of LLD on low ing as much as 20 [14] to 30 mm [15], while others
back pain (LBP) [4], osteoarthritis (OA) of the hip [5], define a significant discrepancy in terms of functional
stress fractures [6], aseptic loosening of hip prostheses outcomes [16]. Perhaps the most controversial muscu-
[7], standing balance [8], forces transmitted through the loskeletal disorder associated with LLD is LBP. Some
hip [9], running economy [10] and associated running authors have found a definite association between LLD
injuries [11]. and LBP [4,17 –21] while others found none [22 –25].
LLD can be subdivided into two etiological groups: a Regarding standing posture, several authors have found
structural LLD (SLLD) defined as those associated an association between LLD and scoliosis [26 –28],
with a shortening of bony structures, and a functional while one study has found the association less clear
[29]. Regarding gait, several authors have found that
* Tel.: +1-272-6850. relatively small (20 –30 mm), LLD created significant
E-mail address: [email protected] (B. Gurney). changes in gait such as increased ground reaction forces

0966-6362/02/$ - see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 9 6 6 - 6 3 6 2 ( 0 1 ) 0 0 1 4 8 - 5
196 B. Gurney / Gait and Posture 15 (2002) 195–206

(GRF) [30,31] increased energy consumption [32], and Table 2

LLD necessary to affect patients–use of subjective criteria
increased lower extremity kinetic energy [33] while
other authors have found that these parameters remain Author Magnitude of Problem/outcome measure
relatively unchanged until much larger LLD (60 mm) LLD (mm)
are realized [34,35]. Tables 1 and 2 provide an overview
of the results of several studies that measured the Goftona 11 (mean) LBP/pain levels
Helliwella 20 (minimum) LBP/pain levels
magnitude of LLD necessary to affect subjects using Rossvolla 32 (mean) LBP/pain levels
both objective (Table 1) and subjective (Table 2) Tjernstroma 60 (mean) LBP/pain levels
criteria. Goela 20 (case study) Meralgia paresthetica
This review was written to consolidate information Friberga 5 LBP/pain levels
regarding the etiology, measurement, complications and Messiera 6.4 Iliotibial band syndrome, shin
splints, plantar fasciitis
treatment of LLD. In the conclusions, an attempt is
made to generalize about the amount of LLD necessary a
Actual LLD.
to create problems in specific populations.
2. Etiology of LLD
Table 1
LLD necessary to affect patients–use of objective criteria
2.1. Etiology of structural LLD
Author Magnitude of Problem/outcome measure
LLD (mm) SLLD, also known as true LLD is defined as differ-
ences in leg length resulting from inequalities in bony
Giles (1982)a 9 mm Pain/arthritic changes
(minimum)
structure. The etiology of SLLD may be congenital or
Giles (1981)a 9 mm Lumbosacral facet joint acquired. Of the congenital causes, the most common
(minimum) angles include congenital dislocation of the hip, and congenital
Youngb 15 mm Pelvic torsion hemiatrophy or hemihypertrophy with skeletal involve-
(minimum) ment. Acquired causes can be as a result of infections,
Cummingsb 6.3 mm Pelvic tilt
(minimum)
paralysis, tumors, surgical procedures such as pros-
Spechta 6 mm Scoliosis or altered lordosis thetic hip replacement, and mechanical such as slipped
(minimum) capital femoral epiphysis.
Papaioannoua \22 mm Scoliosis
Maharb 10 mm Postural sway, shift of mean
(minimum) center of pressure
2.2. Etiology of functional LLD
b
Brand 35 mm Altered forces at the hip
(minimum) Functional, or apparent LLD is a result of muscle
Schuita 10.4 mm Altered GRF (tightens/weakens) or joint tightness across any joint in
(mean) the lower extremity or spine. Some of the more com-
Bhavea 49 mm Asymmetrical stance times,
(mean) GRF
mon causes can be pronation or supination of one foot
Blakea 3.2 mm Increased rear foot eversion in relation to the other, hip abduction/adduction tight-
(minimum) during midstance ness/contracture, knee hyperextension due to quadri-
Kaufmana \20 mm Gait asymmetry ceps femoris weakness, and lumbar scoliosis.
Vinkb 40 mm Increased low back EMG
activity
Delacerdaa 26.7 mm Increased kinetic energy
(case study) during walking, increased 3. Measurement of LLD
oxygen consumption
submaximal during running
Songa 5.5% Increased mechanical work,
3.1. Radiography and other imaging techniques
greater vertical center of mass
displacement with gait Radiography has long been considered to be the gold
Liua \23 mm Gait asymmetry standard for measuring LLD [36], although radio-
Gurneyb 20 mm Increased oxygen
graphic techniques vary and they are not without prob-
consumption in gait
Gurneyb 30 mm Increased HR, minute
lems. There are three methods used that utilize
ventilation, EMG activity of radiographs to measure LLD.
quad on longer leg The first is the ‘orthoroentgenogram’, which involves
Gurneyb 40 mm Increased EMG activity of a single exposure of the legs including hips and ankles.
lower extremity, plantar flexor It has the advantage of requiring only one exposure,
on shorter leg
but is subject to distortion by parallax error [37].
a
Actual LLD. The second method is the ‘scanogram’ which uses
b
Artificially induced LLD. three exposures, one for the hip, knee and ankle, and
 B. Gurney / Gait and Posture 15 (2002) 195–206 197

negates the magnification error, but increases the time, FLLD may be involved. Gross et al. [44] used the
cost, and radiation exposure of the patient. indirect method with a pelvic leveling device and ob-
The third method is the ‘computerized digital radio- tained good intratester reliability (ICC= 0.84), fair in-
graph’. This technique minimizes radiation exposure, tertester reliability (ICC= 0.77), and validity against
reduces mathematical error and is accurate even in the radiographs ranging from 0.55 to 0.76. On the other
presence of angular deformity [38]. All radiographic hand, Friberg et al. [45] reported that the indirect
techniques measure from some landmark on the proxi- method is an inaccurate and imprecise method, with a
mal femur/pelvis to some landmark on the ankle, and 7.5-mm mean difference compared to radiographs, and
do not account for the contribution of the foot to limb a 1.5-mm intratester error. Clarke [36] tested the indi-
length. rect method and reported that two testers were within 5
Computerized tomography (CT), three-dimensional mm of the radiographic leg length in only 16 of 60
(3-D) ultrasonography (US), and magnetic resonance subjects. In addition, Mann et al. [46] showed poor
imaging (MRI) are also used to determine leg length. reliability of determination of the iliac crest height,
CT has been shown to have sensitivity better than 1 which is a requisite for the indirect method.
mm and has shown good reproducibility and little The literature testing the TMM has shown similar
exposure to radiation. Although it is considered more conflicting results. Friberg et al. [45] reported that the
precise and trustworthy than using radiographs [39], ASIS to medial malleolus measurement is an inaccurate
especially when a flexion deformity is present in the and imprecise method; with a mean difference in LLD
knee [40], it is more expensive. Even though conven- measure of 8.6 mm compared to radiographs, and a 1.1
tional US has been proven to be inferior to CT for mm intratester mean error. The authors of this study
determination of leg length, 3-D US has shown to be an used a single measurement of the ASIS to the medial
accurate single step determiner of LLD without ioniz- malleolus. In agreement with this finding, Beattie et al.
ing radiation [41], with a standard deviation for repro- [43] found validity estimates of LLD to have an intra-
ducibility of leg length measurement of 1.6 mm. MRI class correlation coefficient (ICC) of r= 0.683 when
also has the advantage of no radiographic risks, but has utilizing a single measure using the ASIS to medial
been shown to have less reproducibility and accuracy malleolus measure. However, when the average of two
when compared to CT or US [42]. measures was taken from ASIS to the medial malleolus,
the validity of the TMM was good, the ICC with
3.2. Clinical methods radiographs was r= 0.793. In addition, Gogia and
Braatz [47] reported an ICC with radiographs of r=
In general, although the above imaging techniques 0.98 and a between- and within-tester reliability of
are considered to be the most accurate method for r=0.98. Hoyle et al. [48], noted an intertester reliability
determining LLD, they are costly, time consuming, of r= 0.98 and an intratester reliability ranging from
and, in the case of radiographs and CT, the patient is r=0.89 to 0.95 for the ASIS to medial malleolus
exposed to radiation. As a result, alternative clinical measurement.
methods have been developed. Two methods have In conclusion, radiographs or other imaging tech-
emerged over the years: (a) an ‘indirect method’ done in niques should be used when accuracy is critical. There
standing using lift blocks under the short leg and is still disagreement regarding the validity and reliabil-
visually examining the level pelvis [2], and (b) a ‘direct ity of both the indirect methods using a pelvic leveling
method’ done in supine measuring the distance of fixed device and TMM. The average of two measures be-
bony landmarks with a measuring tape. Two commonly tween the ASIS and the medial malleolus appears to
used tape measure methods (TMM) include measuring have acceptable validity and reliability when used as a
the distance between (a) the anterior inferior iliac spine screening tool.
(ASIS) and the lateral malleolus [2] and (b) the ASIS
and the medial malleolus [43]. There is disagreement as
to the validity and reliability of these methods. 4. Treatment
The indirect or standing method incorporates the
contributions of the foot and ankle whereas the TTM Prediction of LLD at skeletal maturity is an impor-
does not. Woerman and Binder-MacLeod [2] compared tant prerequisite for determining the necessary treat-
the indirect method and both TMM against radio- ment to equalize leg length. In order to determine this,
graphs and reported that the indirect method was more future growth potential must be estimated. The study of
accurate and precise than either of the direct methods. growth as it pertains to LLD involves the relationship
Of the two direct methods, they reported the ASIS to between leg length, maturity or skeletal age, and
lateral malleolus measure to be superior to the ASIS to chronological age. The three relationships must be ex-
medial malleolus measure. The authors advocated the amined individually to help predict future growth and
use of the indirect method, especially in cases where therefore, intervention [49].
198 B. Gurney / Gait and Posture 15 (2002) 195–206

Treatment of LLD ranges from shoe inserts to vari- dures is sometimes performed. The maximum shorten-
ous surgical techniques including limb lengthening and ing that can take place in the femur is about 50–60 mm
shortening, and epiphysiodesis. There is disagreement before serious muscle function is compromised, in the
regarding the correct treatment in regards to magnitude tibia, 20– 30 mm [55].
of LLD. Reid and Smith [15] suggest dividing LLD into Limb lengthening is generally reserved for LLD
three categories, mild (0– 30 mm), moderate (30–60 greater than 40–50 mm. This usually involves cortical
mm), and severe (\60 mm), where mild cases should osteotomy (corticotomy) followed by the extremity be-
either go untreated or treated non-surgically, moderate ing fitted with an external fixation device that applies
cases should be dealt with on a case by case basis and continuous longitudinal distraction across the os-
some should be dealt with surgically, and severe cases teotomy site. The lengthening rate is usually about 1
should be corrected surgically. Moseley suggests a simi- mm per day to optimize proper osteoblastic activity.
lar breakdown: 0–20 mm requiring no treatment, 20– One of the more commonly used devices is the Ilizarov
60 mm requiring a shoe lift, epiphysiodesis, or system, which is a cage that surrounds the surgical site
shortening, 60–200 mm requiring lengthening which consisting of rings and small wires that apply the tensile
may or may not be combined with other procedures, force. Although significant gains can be made in bone
and \200 mm prosthetic fitting [49]. length (over 150 mm), the procedure has a host of
complications including fractures above and below the
4.1. Non-surgical inter6ention fixator, bending and fracture of the regenerate bone,
pin-tract infections, joint stiffness, and cyst formation
The most common treatment for mild LLD is the use in the lengthened bones [56]. Other possible complica-
of shoe lifts, which consists of either a shoe insert or tions include sciatic nerve compression due to piri-
building up the sole of the shoe on the shorter leg. In formis tightening [57], subsequent severe physeal
general, up to a 20 mm of correction can be made with growth retardation [58], and popliteal artery injury [59].
an insert, while further corrections should be done on
the sole of the shoe. Reid and Smith [15] state that full
correction with shoe inserts is possible with up to a 10 5. Complications of LLD
mm LLD. They suggest that discrepancies between 10
and 30 mm should be partially corrected with shoe lifts, There are a plethora of studies identifying complica-
but that correction need only be made to 10 mm LLD. tions associated with LLD, but there is little agreement
Moseley suggests that shoe lifts can be used with LLD regarding how much LLD is necessary to cause them.
up to 60 mm [49]. One source suggests that a LLD of over 20 mm is
unacceptable to the patient and can lead to gait anoma-
4.2. Surgical inter6ention lies and spinal deformities [3], while another source
stated that many LLDs up to 30 mm can go untreated
In children with LLD of between 20 and 50 mm, [15], while still another author stating that, with LLD
some authors suggest epiphysiodesis. Epiphysiodesis is up to 50 mm, disability is negligible [60]. Gross [14]
the surgical interruption of the epiphyseal plate on the found that only 50% of patients with a LLD of between
longer leg [50]. The surgery involves the use of a curette 20 and 30 mm felt they were unbalanced, and only 45%
to remove part of the growth plate resulting in a of patients were willing to wear a lift with LLD up to
permanent premature closure, or fusion of the 30 mm.
epiphysis. It appears that the age of onset may be one determin-
Physeal stapling is the placement of staples across the ing factor. Children can compensate for LLD more
epiphysis to temporarily arrest bone growth. Several readily than adults who have experienced a sudden
studies have reported good results with this procedure onset of LLD [61]. Another factor might be the func-
in adolescents [51] as well as children [52] in manage- tional activities of the individual. Athletes may have
ment of knee deformities. There are several complica- symptoms with significantly smaller LLD than non-ath-
tions associated with this procedure including letes [62]. While Siffert [63] reported that a LLD of
permanent arrest of the epiphyseal plate [53] resulting 10– 25 mm is rarely symptomatic in the general popula-
in over correction [49] as well as rebound longitudinal tion, Friberg [6] found that Finnish Army conscripts
growth after removal of staples [54]. Because of the with as little as 10 mm LLD involved in extensive
tendency of permanent arrest of epiphyseal plate training have a greater incidence of stress fractures than
growth, several authors have suggested that epiphyseal controls. Subotnick [1] proposed that a 1/4 in. LLD in
stapling should be considered a permanent form of the athlete is as pathologically important as a 3/4 in.
growth arrest [54]. LLD in the non-athlete.
In patients with skeletal maturity who can accept the Complications associated with LLD can be divided
loss of stature, limb shortening by bone resection proce- into two categories: (a) functional limitations such as
 B. Gurney / Gait and Posture 15 (2002) 195–206 199

gait and balance problems, and (b) associated muscu- and those of Mahar et al. [69]. In a rebuttal paper,
loskeletal disorders such as LBP or stress fractures. Kirby et al. [70] stated that the differences in the
findings between their study and Murrel’s might be due
5.1. Functional limitations associated with LLD to the small amount of average LLD used by Murrel,
as well as differences in foot position during stance.
LLD has been shown to affect several functional
activities such as standing posture, standing balance, 5.1.2. The role of LLD on walking
walking, and running. However, the conflicting infor- Studies on the effects of LLD on walking have found
mation in the literature, combined with the differences gait asymmetries manifest throughout the kinetic chain
in the methodology between studies, make it difficult to [9,30,71]. In general, the individual with a LLD must
make many generalizations. step down onto the short limb and vault over the long
limb resulting in an increase in vertical displacement of
5.1.1. The role of LLD on standing posture/balance the center of mass, and hence an increase in energy
Compensation for LLD in standing can occur in consumption. Gait characteristics with LLD include
many different ways. The longer leg is often compen- decreased stance time and step length on the shorter
sated by pronation of the foot on the longer leg [64]. In leg, decreased walking velocity, and increased walking
addition, supination and/or plantar flexion of the foot cadence [31,72]. Various compensatory mechanisms can
of the shorter leg [65] can be used as a compensatory occur to lengthen the shorter limb including increasing
mechanism. The knee and hip can also compensate by downward pelvic obliquity, increasing knee extension in
extension of the shorter limb and/or flexion of the midstance, vaulting, toe walking, or any combination
longer limb [66]. If the leg is left uncompensated, the of these [30]. In addition, an individual may shorten the
anterior and posterior iliac spines are lower on the side longer leg by increasing pelvic obliquity, circumduction,
of the short leg [67], which, in turn, may result in a increasing hip and/or knee flexion (steppage gait), in-
sacral base unleveling [68] and/or scoliosis [26]. In creasing ankle dorsiflexion, or any combination of these
electromyographic studies, it has been shown that rela- [30]. In addition, Blake and Ferguson [73] found that
tively small leg length discrepancies of 10– 20 mm can there was a significant difference in calcaneal position
lead to a large increase in muscle activity of several of the short and long leg during midstance of gait, with
muscle groups, making it impossible to maintain a the longer side being more everted by 3°. They con-
complete resting position while standing [27]. In addi- cluded that eversion may be a compensatory shortening
tion, a greater amount of pressure is transmitted mechanism, as it has been postulated that eversion of
through the hip of the longer leg due to both a decrease the calcaneus can result in functional shortening of the
in the area of contact of the femoral head on the leg. The patient adopts these compensatory mecha-
acetabulum, as well as an increase in tone of the hip nisms to minimize the displacement of the body center
abductors secondary to an increased distance between of mass during gait, thereby reducing energy expendi-
origin and insertion [28]. To compound this decrease in ture [30].
contact area, there may be an increase in axial loading D’Amico et al. [72] analyzed the gait of 17 patients
through the long leg. Mahar et al. [69] reported that a with an electrodynographic analyzer both before and
10 mm lift on one leg resulted in a significant shift in after correction with a heel lift. They found that prior
the mediolateral position of the center of pressure to correction, the patient’s longer limb demonstrated an
towards the longer leg. average cadence of 48.2 steps/min (1.25 s/step average)
The effect of LLD on standing balance has been and the patient’s shorter limb 52.3 steps/min (1.15
studied. It appears that there is a difference between s/step average). After correction, the patient’s longer
artificially induced LLD and true LLD. In one study, limb demonstrated a cadence of 44.0 steps/min (1.36
the authors measured postural sway changes in stand- s/step average) and the patient’s shorter limb 45.0
ing associated with artificially induced LLD. They steps/min (1.33 s/step average). In addition, they found
found there was a statistically significant increase in that the shorter limb underwent compensatory supina-
postural sway with each increase in induced LLD. The tion and the longer limb compensatory pronation. Cor-
authors concluded that minor LLD might be biome- rection resulted in a reduction in both the pronatory
chanically significant [69]. forces and the supinatory forces during gait.
On the other hand, another study [8] found conflict- Bandy and Sinning [74] examined range of motion
ing findings using persons with true LLD. They found (ROM), duration, and angular velocities in the hip,
no statistical difference in postural sway between sub- knee and ankle during gait in four males with LLD
jects with LLD and controls. The authors concluded ranging from 4.8 to 9.5 mm with and without corrective
that the long-term adaptation by the neuromuscular heel lifts. They found that correction did not signifi-
system found in their subjects (age range, 20– 32 years cantly affect any of the above parameters, but noted
old) accounted for the difference between their findings that the heel lift did cause more symmetrical movement
200 B. Gurney / Gait and Posture 15 (2002) 195–206

for the maximum angle of hip extension and ROM of and long limbs. The authors suggested that the body is
the swing plantar flexion phase on the ankle. able to compensate for minor LLD of up to 20 mm.
Vertical GRF during locomotion have been exam- Mechanical work and walking economy have been
ined by several authors. While a normal gait has been found to be correlated [79]. Song et al. [71] measured
shown to have equal GRF for both limbs [75], several mechanical work of children with true LLD during
authors have found that LLD creates an asymmetry in gait. They found that the children demonstrated several
GRF, with a larger GRF consistently found on the compensations including toe walking, vaulting, circum-
longer leg [30,31]. Since a relationship between GRF duction, and increased flexion of the longer limb. In
and economy during locomotion has been established addition, they noted the long limb performed more
[76], GRF values might also lend some insight into mechanical work than the short limb. The average LLD
energy expenditure during gait. for patients who had no observable compensatory strat-
Bhave et al. [31] used a three parameter GRF analy- egy was 16.4 mm.
sis using a force plate system as their outcome measure. In a single case study, Delacerda and Wikoff [33]
They found significant differences in GRF between measured the lower extremity kinetic energy of a
subjects with true LLD and controls, as well as a woman with a true LLD of 28.7 mm during gait, both
significant difference in GRF between the longer and with and without a corrective lift. They found that the
shorter leg in patients with LLD, with the longer leg addition of the lift decreased the overall kinetic energy
having the greater GRF. Their study included a post- expended by both legs during gait. Gurney et al. [32]
test of the same patients after corrective lengthening used an artificial LLD (aLLD) on healthy older persons
surgery. They reported that after the leg lengths were (ages 55–86) during gait and found that subjects had a
corrected to within 10 mm, there is no significant significantly greater energy consumption using VO2 at
difference in GRF between legs. In addition, correction 20, 30 and 40 mm aLLD compared to no LLD, signifi-
of LLD equalized the stance times between legs from a cantly greater HR at 30 and 40 mm aLLD, significantly
12% difference before surgery to a 2.4% difference greater minute ventilation at 30 and 40 mm aLLD, and
after. reported a higher rate of perceived exertion at 20, 30
Kaufman et al. [30] used five GRF parameters com- and 40 mm aLLD compared to no aLLD. Based on the
bined into an asymmetry index as their outcome mea- findings in studies using GRF, forces at the hip, me-
sure, and reported that 45–70% of the variation in gait chanical work, and kinetic energy, and oxygen con-
asymmetry was explained by the variance in limb sumption, it would appear that a LLD as small as 20
length. They found that a LLD of \ 20 mm resulted in mm can consistently affect the function of gait.
a gait asymmetry significantly greater than that ob- In apparent conflict with the above findings, how-
served in the normal population. ever, Phelps et al. [34] reported that young adults who
Liu et al. [77] performed a similar analysis to Kauf- had as much as 60 mm of true LLD did not have a
man et al. but used GRF in conjunction with several greater oxygen consumption or oxygen cost than nor-
positional parameters to determine what they called a mal adult controls during walking. Richter [35] found
symmetry index (SI). In contrast to Kaufman et al., that only persons with true LLD of E 40 mm had
their results suggested that GRF alone and LLD did significant differences in heart rate. He also performed
not correlate well (they did not report the correlation a repeated measures study and found that subjects
coefficient), and that differences in hip flexion at initial ambulating with as much as a 60 mm aLLD did not
contact and knee minimal flexion during midstance have significantly higher heart rates than with no leg
mainly determined the alterations of subjects SI influ- length difference.
enced by LLD. They concluded that acceptable gait There is little research addressing the effects of LLD
symmetries were evident in individuals with mean val- on muscle activity during gait using electromyography
ues of LLD up to 23.3 mm. (EMG) as a measure. Vink and Huson [80] artificially
Brand and Yack [9] examined the relationship be- induced LLD of 10, 20, 30 and 40 mm and examined
tween LLD and forces at the hip joint during walking. back extensor muscle activity with surface EMG
Unlike the previous GRF studies, these authors artifi- (sEMG). They found a significant increase in sEMG
cially induced LLD of 23, 35, and 65 mm in normal activity of the back extensor muscles during heel strike
individuals. They found that the 35 and 65 mm lifts of the raised limb when LLD was 30 mm or greater.
increased mean peak intersegmental resultant hip forces The authors suggested that the increase in sEMG was
of the short leg by 2 and 12%, respectively, and de- due to an exaggerated trunk flexion induced by an
creased the forces of the long leg by 6 and 12%, increased deceleration of the pelvis during heel strike of
respectively. The 23 mm lift produced no measurable the raised limb. Gurney et al. [32] found that there was
changes. In agreement with these findings, Goel et al. a significant overall increase in muscle recruitment (the
[78] artificially induced LLD of 12.5 mm and noted no sum of quadriceps femoris, gastrocnemius, gluteus max-
significant differences in joint moments between short imus and gluteus medius bilaterally) with a 40 mm
 B. Gurney / Gait and Posture 15 (2002) 195–206 201

aLLD compared to no LLD, but not with a 20 or 30 In apparent conflict with this generalization, how-
mm aLLD. In addition, they found that the quadriceps ever, one study performed by Reid, Smith and Raso
on the longer leg showed significant increases in activity (1982, unpublished data) found no statistically signifi-
only with a 40 mm aLLD, and the plantar flexor on the cant difference with regard to the amount of oxygen
shorter leg showed a significant increase in activity with required to run, even with a 30 mm aLLD.
both a 30 and 40 mm aLLD.
5.2. Associated musculoskeletal disorders
5.1.3. The role of LLD on running
The mechanics of running differs considerably from LLD has been implicated in a variety of disorders
those of walking, and the effect that LLD has on including LBP, scoliosis, pelvic and sacral malalign-
running reflects those differences. The vertical oscilla- ments, arthritis of the spine, hip pain and OA, lower
tion is greater in running compared to walking, and extremity stress fractures, aseptic loosening of the pros-
there is no double support in running, so the weight is thetic hip, trochanteric bursitis, myofascial pain syn-
not shared between the two limbs. In addition, the drome (MPS) of the peroneus longus, patellar apicitis,
stance phase of walking is about 60% of the total gait and meralgia paresthetica.
cycle as compared to 30% with running [81]. These
differences lead to stresses on the lower extremity that 5.2.1. Low back pain
are three times that of walking [12]. It has been sug- Perhaps, the most equivocal pathological condition
gested that biomechanical abnormalities due to LLD associated LLP is LBP. LBP is a general term that
are three times more significant in running compared to includes pain in the lumbar spine, the lumbosacral
walking [1]. junction, and the sacroiliac joint (SI). LLD appears to
The findings of the literature vary regarding the affect the lumbar spine, at least in part, by creating a
effects of LLD on running dynamics. Studies have lumbar scoliosis. It has been shown that LLD leads to
addressed parameters including rear foot position and pelvic obliquity on the frontal plane and a scoliosis that
energy consumption. Blake and Ferguson [73] exam- has been described as compensatory, non-structural,
ined rear foot position of persons with true LLD during and non-progressive [26]. Friberg [17] described the
running. They found that there are significant differ- scoliosis as a lumbar convexity toward the short leg
ences between the calcaneus-to-vertical angle of the side with a concomitant vertebral axial rotation. He
short and long legs in both early midstance and goes on to say that the curve is produced solely by the
midstance, with the longer leg exhibiting greater cal- intervertebral joints, and the subsequent wedging on
caneal eversion. Overall, the difference in calcaneal the intervertebral discs in combination with the axial
position was greater in running when compared to rotation are predisposing factors to LBP. Young et al.
walking. Unfortunately, the authors did not report the [84] examined the immediate effects of a simulated
magnitude of the LLD in their subjects except to say LLD of a magnitude necessary to create a lateral tilt of
they were over 1/8 in. (3.2 mm). In contrast to these the pelvis of 1.2° and above on pelvic torsion and trunk
findings, Bloedel and Hauger [82] reported no signifi- flexion. They found the innominate contralateral to the
cant difference in the amount of maximum calcaneal lift rotated anteriorly compared to the ipsilateral side
inversion or eversion between the long and short limbs. and that lateral flexion of the trunk increased toward
The range of true LLD in their subjects was from 12.7 the side of the lift. Giles and Taylor [85] found that
to 19.0 mm. However, the authors did not control for patients with LLD demonstrated abnormal radiologic
shoe type, which has been shown to influence the rear findings compared to controls including wedging of the
foot during initial contact to midstance [83]. fifth lumbar vertebra, concavities of the vertebral end-
Delacerda and McCrory [10] found that a corrective plates in the lumbar spine, and traction spurs and
lift resulted in a reduction of oxygen consumption in a osteophytes of the vertebral bodies. Morscher [28] re-
30-year-old runner with a 28.6 mm true LLD when ported that the asymmetrical loading forces acting on
running at a sub-maximal rate compared to running the spine secondary to scoliosis caused by LLD result
without a lift. In a similar study, Kern (1995, unpub- in early degeneration affecting both the intervertebral
lished data) examined oxygen consumption in runners disc spaces in the form of osetophyte formation as well
with true LLD of 3– 11 mm. He found no differences in as arthosis of the facet joints on the concavity of the
oxygen consumption of subjects with and without a scoliosis. Hoikka et al. [29] found that the relationship
corrective shoe lift. In fact, he noted that the running between LLD and scoliosis was not as clear. They
economy actually worsened in some subjects with the reported that although LLD correlated well (r=0.843)
use of the orthotic correction. In conclusion, based on with pelvic tilt (on the frontal plane), it only correlated
studies that reported LLD magnitudes, it appears that moderately with sacral tilt (r=0.639), and poorly with
the LLD needs to in excess of 19 mm before running lumbar scoliosis (r= 0.338). They proposed that the
parameters are affected. body compensates for asymmetry associated with LLD
202 B. Gurney / Gait and Posture 15 (2002) 195–206

progressively up the kinetic chain. They concluded that Giles and Taylor [4] treated 50 LBP patients with LLD
although there may be an association between LLD with shoe lifts. After a 4-month follow-up, their results
and LBP, LLD might not be the cause of LBP. showed less working days lost to LBP, a marked de-
Giles [86] examined patients with LLD radiographi- crease in subject-reported symptoms and numbers of
cally and found that the lumbosacral facet joint angles LBP episodes, and an increase in overall ROM. In a
on the short side were smaller with the horizontal clinical study by Gofton, ten patients with LLD and
compared to the controls. He postulated that the asym- LBP who were treated with a shoe lift on the short side
metry of joint angles could predispose patients to os- experienced major or complete pain relief that lasted
teoarthritic changes at the lumbosacral joints. upon follow-up ranging from 3 to 11 years [19]. In a
It has been demonstrated that LLD is associated with similar clinical study by Helliwell [20], 18 patients with
sacroiliac malalignment. Greenman [68] stated that LBP and a LLD of 20 mm or more were treated with
LLD can result in sacral tilt, and that patients with an corrective shoe lifts. Upon follow-ups of at least 3
unlevel sacral base were twice as likely to complain of months, 44% experienced complete pain relief, and 45%
low back syndromes as patients without an unlevel had substantial or moderate pain relief. Similarly,
sacral base. He concluded that a sacral base unleveling Friberg [17] found that, of the 211 persons with LBP
of greater than 4 mm as determined by X-ray should be treated with shoe lifts, 157 were symptom-free after a
considered clinically significant. Schuit et al. [87] found mean follow-up of 18 months. In a study by Rossvoll et
that nine of the 14 subjects with LLD had a SI al. [21], 22 LBP patients with an average of 32 mm
malalignment of some kind. Pitkin and Pheasant [88] LLD were surgically managed with shortening os-
described an innominate bone rotation of between 3 teotomy. After surgery, their average LLD was 4.3 mm,
and 19° in patients with LLD. Cummings et al. [89] and their LBP was significantly reduced. Finally, Tjern-
reported the same relationship and described it as a strom and Rehnberg [90] performed 100 leg-lengthening
posterior innominate bone rotation on the side of the procedures. Before lengthening, 18 of these patients
lengthened limb and an anterior rotation on the side of experienced LBP, whereas, after an average 6-year post-
the shorter limb. The authors also suggested that in- surgical follow-up, six patients experienced LBP.
nominate rotation could limit movement strategies or On the other hand, several investigators have found
patterns during activities that involve movement of the no relationship between LBP and LLD. Hellsing [22]
pelvis, that may result in asymmetrical loading of the SI examined over 600 military trainees over a period of 4
ligaments, predisposing them to chronic strain. In addi- years and found no correlation between LLD and back
tion, they proposed that a constant asymmetrical load- pain or pain provocating tests. Similarly, Nadler et al.
ing on the SI joint surfaces might lead to degenerative [23] showed in a prospective study on 257 college
changes of the joint. athletes that LLD was not associated with future LBP
There have been many studies showing a positive treatments. Soukka et al. [24] found no association
relationship between LLD and LBP. In the introduc- between mild LLD measured radiographically and LBP
tion of an article by Giles and Taylor [4], the authors in 247 men and women. Yrjonen et al. [25] found LLD
combined five studies with a total population of 1806 and lumbar scoliosis correlated poorly with LBP in 96
subjects. They chose to look only at studies that used patients with Perthes’ disease.
radiographic measurements of LLD. They reported
that about 7% of the subjects with no history of LBP 5.2.2. Hip pain
have a LLD of 10 mm or more, while four of the five The literature appears less equivocal regarding hip
studies showed 13– 22% of subjects who sought medical symptomology. Brunet et al. [11] surveyed 1493 male
treatment for their LBP have LLD. In their own study, and female runners and found that hip pain is over
Giles and Taylor tested 217 persons. Of these, 8% who twice as common a complaint in persons with a self-re-
had no complaints of LBP had a LLD of 10 mm or ported LLD compared with persons with no LLD.
more, while 18.3% of those with LBP had a LLD of 10 Friberg [17] reported that of 254 patients with LLD
mm or more. Friberg [17] compared 653 patients with complaining of chronic hip pain, 226 have pain on the
LBP with 359 controls with radiography and found the longer extremity. Twenty-seven of these subjects had
relative prevalence of leg length inequality of \ 5mm severe idiopathic hip arthrosis, with the arthrosis occur-
was 1.7 times greater in the LBP group compared to the ring on the longer leg in 24 of the 27 subjects. Friberg
control group. The prevalence grew to 5.3 times greater treated 79 of these hip patients with shoe lifts, 56 of
with 15 mm LLD. In a population of 132 persons with which became symptom-free after treatment. Gofton
LLD of 10 mm or greater who were experiencing LBP and Trueman [5] had similar results. They examined 67
and radiating leg pain, 104 had pain radiating into the patients with hip OA, of which 62 were idiopathic. Of
shorter leg (P=0.02) [18]. these, 36 patients demonstrated significant LLD. All 36
If LLD does lead to LBP, then treatments aimed at patients had superolateral type OA, and 29 of the 36
equalizing the LLD should help with the symptoms. patients had the arthritic findings in the longer leg.
 B. Gurney / Gait and Posture 15 (2002) 195–206 203

Since OA is generally considered to cause shortening of no LLD experienced stress fractures, 46.2% of those
the involved leg, the authors dismissed the possibility with a 10–14 mm LLD had stress fractures, and 66.7%
that the OA is the cause of the LLD noted in these of those with LLD of 15–20 mm had stress fractures.
patients. They suggested that a longer leg might be a He also reported that stress fractures occurring in the
predisposing factor to hip OA. tibia, metatarsals, and femur were found in the long leg
This relationship between hip OA and the long leg in 73% of the time. The increased incidence of stress
LLD is consistent with the calculations of Krakovits fractures on the longer limb seen by Friberg appears
[91] who mathematically modeled a relationship be- consistent with the greater forces emitted through the
tween LLD and reduction of weight-bearing area of the longer leg described under hip pain.
femoral head. Based on his formula, a 10 mm increase
in leg length would result in a reduction of weight-bear- 5.2.4. Other pathology
ing area of 5%, and an increase in leg length of 50 mm In addition to the above problems, there appears to
would cause a reduction of contact area of 25.1%. be an association between LLD and several other lower
The work of Morscher [28] also supports these find- extremity problems. Swezey [94] found a relationship
ings. He found a greater amount of pressure transmit- between LLD and trochanteric bursitis. He found a
ted through the hip of the longer leg due to the decrease LLD of 25.4 mm or greater in 13 of 20 patients
in area of contact of the femoral head on the acetabu- diagnosed with trochanteric bursitis. Kujala et al. [95]
lum, as well as an increase in tone of the hip abductors found significantly more patellar apicitis (PB0.001)
secondary to increase distance between origin and in- and patellofemoral joint incongruencies [96] in patients
sertion. These findings coupled with the finding of with LLD than in controls. Saggini et al. [97] found
several authors that larger GRF are found on the that LLD was related to MPS of the peroneus longus.
longer leg [30,31] suggest the longer leg to be at risk. When they treated MPS patients with shoe lifts, they
Gurney et al. [32] found greater EMG activity on the reported a moderate, significant reduction in pain re-
quadriceps of the longer leg with an aLLD, which ports. Finally, in a single case report, Goel [98] found
could be due to the greater GRF that leg would have to that when he treated a patient with long standing
overcome. Finally, the findings of Visuri et al. [7] seem meralgia paresthetica with a shoe lift in his shorter leg,
to support the greater force transmission through the the pain was resolved.
longer leg. They found that overlengthening of the
prosthetic leg is the most important single variable
predisposing patients to aseptic loosening of the pros- 6. Conclusion
thesis after total hip arthroplasty.
The issue of greater forces through the hip of the There is still controversy regarding the magnitude of
longer leg is still in question, however. Brand and Yack LLD necessary to cause musculoskeletal problems. The
[9] found that when subjects were given aLLD of 35 articles that find associations between LLD and patho-
and 65 mm, they demonstrated decrease intersegmental logical conditions must be interpreted with caution, as
resultant forces through the hip on the longer leg of 6 association does not prove a cause effect relationship,
and 12%, respectively. In addition, Schuit et al. [92] and may just represent a coincident finding.
found that patients with LLD ranging from 4.8 to 22.2 It is clear that studies using persons with long stand-
mm demonstrated a maximum vertical force that was ing true LLD are able to cope with larger LLD than
significantly greater in the short leg before correction those who are subjected to artificial or induced LLD.
with a lift compared to after correction. This is sensible since given enough time most individu-
als would be able to reduce the energy and mechanical
5.2.3. Stress fractures costs of LLD. It also appears that younger persons on
Several studies have shown a higher incidence of the whole are able to adapt to larger LLD than older
stress fractures in the lower extremities in people with persons. Again, this is sensible since it has been shown
LLD. Bennel et al. [93] reported that LLD is a risk that gait patterns differ considerably between old and
factor for stress fractures in female athletes, in that a young individuals [99,100], and that older persons have
significantly greater number of women in the stress greater difficulty in mastering novel motor tasks
fracture group (70%) displayed a LLD compared to the [101,102]. The level of activity of the person also seems
non-stress fracture group (36%). Brunet et al. [11] sur- to play a role. Individuals who are on their feet most of
veyed 1493 male and female runners and noted that the day [17] or who are involved in sports [1] seems to
over twice as many runners with a self-reported LLD be more sensitive to LLD than those who are less
develop stress fractures when compared to runners who active. Finally, studies that use patient pain levels or
reported no LLD. Friberg [6] noted an association satisfaction as an outcome measure seem to differ from
between the quantity of LLD and the incidence of those that use more objective outcome measures (Ta-
stress fractures. While 15.4% of Finnish conscripts with bles 1 and 2).
204 B. Gurney / Gait and Posture 15 (2002) 195–206

The question of whether to treat individuals with [22] Hellsing AL. Leg length inequality. A prospective study of
LLD should be taken on a case by case basis, although young men during their military service. Up J Med Sci
1988;93:245 – 53.
the breakpoint of 20 mm is often used, this number [23] Nadler SF, Wu KD, Galski T, Feinberg JH. Low back pain in
could be considerably higher for younger persons who college athletes. A prospective study correlating lower extremity
are inactive and have had LLD their entire life and overuse or acquired ligamentous laxity with low back pain.
considerably lower for older persons who are active and Spine 1998;23:828 – 33.
have acquired LLD later in life. [24] Soukka A, Alaranta H, Tallroth K, Heliovaara M. Leg-length
inequality in people of working age. The association between
mild inequality and low-back pain is questionable. Spine
1991;16:429 – 31.
References [25] Yrjonen T, Hoikka V, Poussa M, Osterman K. Leg-length
inequality and low-back pain after Perthes’ Disease: a 28 –47-
[1] Subotnick SI. Limb length discrepancies of the lower extremity year follow-up of 96 patients. J Spinal Disord 1992;5:443 –7.
(the short leg syndrome). J Orthop Sports Phys Ther [26] Papaioannou T, Stokes I, Kenwright J. Scoliosis associated
1981;3:11 – 5. with limb-length inequality. J Bone Joint Surg Am 1982;64:59 –
[2] Woerman AL, Binder-MacLeod SA. Leg length discrepancy 62.
assessment: accuracy and precision in five clinical methods of [27] Taillard W, Morscher E. Beinlängenunterschiede. Basel:
evaluation. J Orthop Sports Phys Ther 1984;5:230 –8. Karger, 1965.
[3] Guichet JM, Spivak JM, Trouilloud P, Grammont PM. Lower [28] Morscher E. Etiology and pathophysiology of leg length dis-
limb-length discrepancy. An epidemiologic study. Clin Orthop crepancies. Prog Orthop Surg 1977;1:9 – 19.
1991;272:235 – 41. [29] Hoikka V, Ylikoski M, Tallroth K. Leg length inequality has
[4] Giles LGF, Taylor JR. Low-back pain associated with leg poor correlation with lumbar scoliosis. A radiological study of
length inequality. Spine 1981;6:510 –21. 100 patients with chronic low back pain. Arch Orthop Trauma
[5] Gofton JP, Trueman GE. Studies in osteoarthritis of the hip: Surg 1989;108:173 – 5.
Part II. Osteoarthritis of the hip and leg-length disparity. Can [30] Kaufman KR, Miller LS, Sutherland DH. Gait asymmetry in
Med Assoc J 1971;104:791 –9. patients with limb-length inequality. J Pediatr Orthop
[6] Friberg O. Leg length asymmetry in stress fractures, a clinical 1996;16:144 – 50.
and radiological study. J Sports Med 1982;22:485 –8. [31] Bhave A, Paley D, Herzenberg JE. Improvement in gait
[7] Visuri T, Lindholm TS, Antti-Poika I, Koskenvou M. The role parameters after lengthening for the treatment of limb-length
of overlength of the leg in aseptic loosening after total hip discrepancy. J Bone Joint Surg Am 1999;81:529 – 34.
arthroplasty. Ital J Orthop Traumatol 1993;19:107 – 11. [32] Gurney AB, Mermier CM, Robergs RA, Gibson AL, Rivero
[8] Murrel P, Cornwall MW, Doucet SK. Leg-length discrepancy: DP. Effects of leg length discrepancy on gait economy and
effect on the amplitude of postural sway. Arch Phys Med
lower extremity muscle activity in older adults. J Bone Joint
Rehabil 1991;72:646 –8.
Surg Am 2001;83:907 – 15.
[9] Brand RA, Yack HJ. Effects of leg length discrepancies on the
[33] Delacerda FG, Wikoff OD. Effect of lower extremity asymme-
forces at the hip joint. Clin Orthop 1996;333:172 – 80.
try on the kinematics of gait. J Orthop Sports Phys Ther
[10] Delacerda FG, McCrory ML. A case report: effect of a leg
1982;3:105 – 7.
length differential on oxygen consumption. J Orthop Sports
[34] Phelps JA, Novacheck TA, Dahl MT. Consequences of leg
Phys Ther 1981;3:17 –20.
length inequality in young adults. In: Annual East Coast Clini-
[11] Brunet ME, Cook SD, Brinker MR, Dickinson JA. A survey of
cal Gait Laboratory Conference, Rochester, Minnesota, May
running injuries in 1505 competitive and recreational runners. J
1993.
Sports Med Phys Fitness 1990;30:307 –15.
[35] Richter J. Beinlängendifferenzen und Gehleistung (ein bei-
[12] Blake RL, Ferguson B. Limb length discrepancies. J Am Podi-
tragzurergometric in der orthopädie) [Differences in the length
atric Med Assoc 1992;82:33 –8.
[13] Etnier JL, Landers DM. Motor performance and motor learn- of legs and walking capacity]. Z Orthop Ihre Grenzgeb
ing as a function of age and fitness. Res Q Exerc Sport 1968;104:538 – 54.
1998;69:136 – 46. [36] Clarke GR. Unequal leg length: an accurate method of detec-
[14] Gross RH. Leg length discrepancy: how much is too much? tion and some clinical results. Rheumatol Phys Med
Orthopedics 1978;1:307 –10. 1972;11:385 – 90.
[15] Reid DC, Smith B. Leg length inequality: a review of etiology [37] Stevens PM. Radiographic distortion of bones: a marker study.
and management. Physiother Can 1984;36:177 – 82. Orthopedics 1989;12:1457 – 63.
[16] Abraham WD. Leg length discrepancy in total hip arthroplasty. [38] Moseley CF. Assessment and prediction in leg-length dis-
Orthop Clin North Am 1992;23:201 – 9. crepancy. Instr Course Lect 1989;38:325 – 30.
[17] Friberg O. Clinical symptoms and biomechanics of lumbar [39] Tokarowski A, Piechota L, Wojciechowski P, Gajos L, Kusz D.
spine and hip joint in leg length inequality. Spine 1983;8:643 – Measurement of lower extremity length using computed tomog-
51. raphy. Chir Narzadow Ruchu Orthop Pol 1995;60:123 –7.
[18] ten Brinke A, van der Aa HE, van der Palen J, Oosterveld F. Is [40] Aaron A, Weinstein D, Thickman D, Eilert R. Comparison of
leg length discrepancy associated with the side of radiating pain orthoroentgenography and computed tomography in the mea-
in patients with a lumbar herniated disc? Spine 1999;24:684 – 6. surement of limb-length discrepancy. J Bone Joint Surg Am
[19] Gofton JP. Persistent low back pain and leg length disparity. J 1992;74:897 – 902.
Rheumatol 1985;12:747 –50. [41] Keppler P, Strecker W, Kinzl L, Simmnacher M, Claes L.
[20] Helliwell M. Leg length inequality and low back pain. Practi- Sonographic imaging of leg geometry. Orthopade
tioner 1985;229:483 –5. 1999;28:1015 – 22.
[21] Rossvoll I, Junk S, Terjesen T. The effect of low back pain on [42] Keppler P, Strecker W, Kinzl L. Analysis of leg geometry –
shortening osteotomy for leg length inequality. Int Orthop standard techniques and normal values. Chirurg 1998;69:1141 –
1992;16:388 – 91. 52.
 B. Gurney / Gait and Posture 15 (2002) 195–206 205

[43] Beattie P, Isaacson K, Riddle DL, Rothstein JM. Validity of [67] McCaw ST, Bates BT. Biomechanical implications of mild leg
derived measurements of leg-length differences obtained by use length inequality. Br J Sports Med 1991;25:10 –3.
of a tape measure. Phys Ther 1990;70:150 – 7. [68] Greenman PE. Lift therapy: use and abuse. J Am Osteopath
[44] Gross MT, Burns CB, Chapman SW, Hudson CJ, Curtis HS, Assoc 1979;79:238 – 50.
Lehmann JR, et al. Reliability and validity of rigid lift and [69] Mahar RK, Kirby RL, MacLeod DA. Simulated leg-length
pelvic leveling device method in assessing functional leg length discrepancy: its effect on mean center-of-pressure position and
inequality. J Orthop Sports Phys Ther 1998;27:285 –94. postural sway. Arch Phys Med Rehabil 1985;66:822 – 4.
[45] Friberg O, Nurminen M, Kouhonen K, Soininen E, Mänttäri [70] Kirby RL, Mahar RK, MacLeod DA, Garner JB. Leg length
T. Accuracy and precision of clinical estimation of leg length discrepancy and postural sway. Arch Phys Med Rehabil
inequality and lumbar scoliosis: comparison of clinical and 1992;73:401 – 2.
radiological measurements. Int Disability Stud 1988;10:49 – 53. [71] Song KM, Halliday SE, Little DG. The effect of limb-length
[46] Mann M, Glasheen-Wray M, Nyberg R. Therapist agreement discrepancy on gait. J Bone Joint Surg Am 1997;79:1690 –8.
for palpation and observation of iliac crest heights. Phys Ther [72] D’Amico JC, Dinowitz HD, Polchaninoff M. Limb length
1984;64:334 – 8. discrepancy. J Am Podiatr Med Assoc 1985;75:639 – 43.
[47] Gogia PP, Braatz JH. Validity and reliability of leg length [73] Blake RL, Ferguson HJ. Correlation between limb length dis-
measurements. J Orthop Sports Phys Ther 1986;8:185 –8. crepancy and asymmetrical rearfoot position. J Am Podiatr
[48] Hoyle DA, Latour M, Bohannon RW. Intraexaminer, interex- Med Assoc 1993;83:625 – 33.
aminer, and interdevice comparability of leg length measure- [74] Bandy WD, Sinning WE. Kinematic effects of heel lift use to
ments obtained with measuring tape and metrecom. J Orthop correct lower limb length differences. J Orthop Sports Phys
Sports Phys Ther 1991;14:263 –8. Ther 1986;7:173 – 9.
[49] Moseley CF. Leg length discrepancy and angular deformity of [75] Cahalan TD, King L, Chao EYS. Symmetry of the vertical
the lower limbs. In: Lovell and Winter’s pediatric orthopaedics, ground reaction force in normals. In: Proceedings of the Eu-
4th ed. Philadelphia: Lippencott-Raven, 1996:877. ropean Symposium on Clinical Gait Analysis, Zurich, Switzer-
[50] Staheli LT. Fundamentals of pediatric orthopedics, 2nd ed. land, 1992, p. 79 – 82.
Philadelphia: Lippencott-Raven, 1998:35 –7. [76] Williams K, Cavanagh PR. Relationship between distance run-
[51] Stevens PM, Maguire M, Dales MD, Robins AJ. Physeal ning mechanics, running economy, and performance. J Appl
Physiol 1987;63:1236 – 45.
stapling for idiopathic genu valgum. J Pediatr Orthop
[77] Liu XC, Fabry G, Molenaers G, Lammens J, Moens P. Kine-
1999;19:645 – 9.
matic and kinetic asymmetry in patients with leg-length dis-
[52] Mielke CH, Stevens PM. Hemiepiphyseal stapling for knee
crepancy. J Pediatr Orthop 1998;18:187 – 9.
deformities in children younger than 10 years: a preliminary
[78] Goel A, Loudon J, Nazare A, Rondinelli R, Hassanein K. Joint
report. J Pediatr Orthop 1996;16:423 –9.
moments in minor limb length discrepancy: a pilot study. Am J
[53] Blackstone BG, Coleman SS. Treatment of lower limb length
Orthop 1997;26:852 – 6.
inequality. Surg Ann 1984;16:259 –96.
[79] Burdett R, Skrinar G, Simon S. Comparison of mechanical
[54] Tachdjain MO. Pediatric orthopedics, vol. 4, 2nd ed. Philadel-
work and metabolic energy consumption during normal gait. J
phia: Saunders, 1990:2888.
Orthop Res 1983;1:63 – 72.
[55] Herzenberg JE. Congenital limb deficiency and limb length
[80] Vink P, Huson A. Lumbar back muscle activity during walking
discrepancy. In: Canale ST, Beaty JH, editors. Operative pedi-
with a leg inequality. Acta Morphol Neerl Scand 1987;25:261 –
atric orthopaedics, 2nd ed. St. Louis: Mosby, 1995:223 – 7.
71.
[56] Guidera KJ, Ogden JA, Ganey TM. In: Dee R, editor. Princi-
[81] Mann R, Baxter D, Lutter L. Running symposium. Foot Ankle
ples of orthopaedic practice, 2nd ed. New York: McGraw-Hill, 1981;1:190 – 224.
1997:757 – 63. [82] Bloedel PK, Hauger B. The effects of limb length discrepancy
[57] Nagi ON, Dhillon MS, Goni V. Does the piriformis compress on subtalar joint kinematics during running. J Orthop Sports
the sciatic nerve during limb length equalization? Singapore Phys Ther 1995;22:60 –4.
Med J 1999;40:749 –51. [83] Clarke TE, Frederick EC, Hamill L. The effects of shoe design
[58] Price CT, Carantzas AC. Sever growth retardation following parameters on rearfoot control in running. Med Sci Sports
limb lengthening: a case report. Iowa Orthop J 1996;16:139 – 46. Exerc 1983;15:376 – 81.
[59] Waldhausen J, Mosca V, Johansen K, Schaller R. Delayed [84] Young RS, Andrew PD, Cummings GS. Effect of simulating
presentation of popliteal artery injury during Ilizarov limb leg length inequality on pelvic torsion and trunk mobility. Gait
lengthening. Orthopedics 1998;2:477 – 8. Posture 2000;11:217 – 23.
[60] D’Aubigne RM, Dubousset J. Surgical correction of large [85] Giles LGF, Taylor JR. Lumbar spine structural changes associ-
length discrepancies in the lower extremity of children and ated with leg length inequality. Spine 1982;7:159 – 62.
adults. An analysis of twenty consecutive cases. J Bone Joint [86] Giles LGF. Lumbosacral facetal ‘joint angles’ associated with
Surg Am 1971;53:411 –30. leg length inequality. Rheumatol Rehabil 1981;20:233 –8.
[61] Dahl MT. Limb length discrepancy. Pediatr Clin North Am [87] Schuit D, McPoil TG, Mulesa P. Incidence of sacroiliac joint
1996;43:849 – 65. malalignment in leg length discrepancies. J Am Podiatr Med
[62] Subotnick SI. The short leg syndrome. Physician Sportsmed Assoc 1989;79:380 – 3.
1975;3(6):61– 3. [88] Pitkin HC, Pheasant HC. Sacrarthrogenetic tetalgia. J Bone
[63] Siffert RS. Lower limb discrepancy. J Bone Joint Surg Am Joint Surg Am 1936;18:365 –74.
1987;69:1100 – 6. [89] Cummings G, Scholz JP, Barnes K. The effect of imposed leg
[64] Langer S. Structural leg shortage: a case report. J Am Podiatr length difference on pelvic bone symmetry. Spine 1993;18:368 –
Assoc 1976;66:38 – 40. 73.
[65] Blustein SM, D’Amico JC. Limb length discrepancy. Identifica- [90] Tjernstrom B, Rehnberg L. Back pain and arthralgia before
tion, clinical significance, and management. J Am Podiatr Med and after lengthening. 75 patients questioned after 6 (1 –11)
Assoc 1985;75:200 –6. years. Acta Orthop Scan 1994;65:328 – 32.
[66] Bolz S, Davies GJ. Leg length differences and correlation with [91] Krakovits G. Statik und dynamik des hueftgelenkes. Zeitschrift
total leg strength. J Orthop Sports Phys Ther 1984;6:123 – 9. für arthropädie 1967;102:418.
206 B. Gurney / Gait and Posture 15 (2002) 195–206

[92] Schuit D, Adrian M, Pidcoe P. Effect of heel lifts on ground [97] Saggini R, Giamberardino MA, Gatteschi L, Vecchiet L. My-
reaction force patterns in subjects with structural leg-length ofascial pain syndrome of the peroneus longus: biomechanical
discrepancies. Phys Ther 1989;69:663 – 70. approach. Clin J Pain 1996;12:30 – 7.
[93] Bennell KL, Malcom SA, Thomas SA, Reid SJ, Brukner PD, [98] Goel A. Meralgia paresthetica secondary to limb length dis-
Ebeling PR, et al. Risk factors for stress fractures in track and crepancy: case report. Arch Phys Med Rehabil 1999;80:348 –9.
field athletes. A twelve-month prospective study. Am J Sports [99] Judge JO, Davis RB. Step length reduction in advanced age: the
Mede 1996;24:810 –8. role of ankle and hip kinetics. J Gerontol A Viol Sci Med Sci
[94] Swezey RL. Pseudo-radiculopathy in subacute trochanteric bur- 1996;51:303 – 12.
sitis of the subgluteus maximus bursa. Arch Phys Med Rehabil [100] Woo J, Lau J, Chan SG, Yuen YK. Age-associated gait
1976;57:387 – 90. changes in the elderly: pathological or physiological? Neuroepi-
[95] Kujala UM, Osterman K, Kvist M, Aalto T, Friberg O. demiology 1995;14:65 – 71.
Factors predisposing to patellar chondropathy and patellar [101] Etnier JL, Lander DM. Motor performance and motor learning
apicitis in athletes. Int Orthop 1986;10:195 – 200. as a function of age and fitness. Res Q Exerc Sport
[96] Kujala UM, Friberg O, Aalto T, Kwist M, Osterman K. Lower 1998;69:136 – 46.
limb asymmetry and patellofemoral joint incongruence in the [102] Bickford P. Motor learning deficits in aged rats are correlated
etiology of knee exertion injuries in athletes. Int J Sports Med with loss of cerebellar noradrenergic function. Brain Res
1987;8:214 – 20. 1993;620:133 – 8.