FOOT & ANKLE INTERNATIONAL
Copyright © 2001 by the American Orthopaedic Foot & Ankle Society, Inc.
Experimental Flatfoot Model: The Contribution of Dynamic Loading
In-Tak Chu, MD; Mark S. Myerson, MD; Meir Nyska, MD*; and Brent G. Parks, MSc
ABSTRACT
The goal of this study was to determine if the application
of muscle forces (simulating the dynamic phase of the
midstance part of gait) had an effect on flatfoot deformi-
ty. We created a flatfoot model in each of seven cadaver
foot specimens by grasping the Achilles, peroneus
longus, peroneus brevis, flexor digitorum longus, and
flexor hallucis longus tendons with soft-tissue vice
clamps connected via wire cables to pneumatic cylinders.
The experiment included four stages: 1) initial static axial
loading; 2) axial loading after 3,000 load cycles (average,
735 N; range, 70 to 1400 N); 3) axial loading after releas-
ing the spring ligament and plantar fascia; and 4) axial
loading after an additional 3,000 load cycles. At each
stage, both static (with axial loading only) and dynamic
(axial loading with tensioning of the tendons to simulate
the muscle forces at midstance) conditions were evaluat-
ed radiographically. No change was observed between
the static and dynamic conditions in the first two phases
of the experiment. After the third phase, changes in the
talar-first metatarsal angle and the height of the medial
cuneiform were noted, particularly in the dynamic condi-
tion. These and additional radiographic changes were
magnified in the fourth phase, but only in the dynamic
condition. We concluded that, to create an effective flat-
foot model, the medial structures, including the spring
ligament and possibly the plantar fascia, must be sev-
ered. Cyclic loading of the foot further increased the arch
flattening, and this effect was magnified by dynamic load-
ing.
INTRODUCTION
A tear of the tibialis posterior tendon leads to acquired
flatfoot in an adult." but repair of this tendon, or recon-
From the Department of Orthopaedic Surgery, The Union Memorial Hospital,
Baltimore, Maryland
'Current affiliation: Department of Orthopaedic Surgery, Foot and Ankle
Service, Hadassah Medical Center, Hebrew University, Jerusalem, Israel
Address all correspondence and reprint requests to; Mark S. Myerson, MD,
c/o Lyn Camire, Editor, Union Memorial Orthopaedics, The Johnston
Professional Building, #400, 3333 North Calvert Street, Baltimore, MD 21218
(410-554-6668; fax: 410-554-2832; e-mail: Iync@helix.org).
220
Baltimore, MD
struction using other tendons, does not consistently
improve the hindfoot valgus. The improvement tends
not to be maintained in long-term follow-up.v" There
has been growing evidence, both clinically and experi-
mentally, that there are secondary capsuloligamentous
structures that stabilize the arch, including, for example,
the spring llqarnent.':" It has been shown that an exper-
imental flatfoot model can be created only if all the
medial structures that stabilize the arch, including the
spring ligament, long and short plantar ligaments, and
the plantar fascia, are cut." In a more recent work,
Kitaoka et al.' evaluated the effect of the posterior tibial
tendon on the arch by creating a dynamic flatfoot
model. They concluded that the tibialis posterior is an
important stabilizer of the arch and that loading the
extrinsic muscles accentuates pes planus. Deland et
al.' created a reproducible severe flatfoot model to
evaluate reconstructive procedures by cutting the
spring ligament, medial portion of calcaneocuboid joint,
long plantar ligament, plantar fascia, superficial anterior
deltoid, capsule of medial subtalar joint, and
interosseus ligament. Recently McCormack et al.10 pro-
duced a dynamic flatfoot model by cutting the spring lig-
ament and utilizing cyclic axial compressive loading.
They reported increased contact area and pressure in
the posterior facet and attributed it to the inclusion of
muscle forces. However, they did not investigate the
effect of static versus dynamic loading on the flatfoot. To
evaluate whether the application of muscle forces has
an effect on the flatfoot deformity, we created a flatfoot
model and then observed the effect of static and
dynamic loads on the foot.
The goal of the current study was to determine if the
application of muscle forces to simulate the dynamic
phase of the midstance of gait had an effect on flatfoot
deformity.
MATERIALS AND METHODS
Preparation of Specimens
Seven fresh-frozen specimens were selected from
cadaver lower extremities after gross visual screening
Foot & Ankle InternationalNol. 22, No. 3/March 2001
for preexisting abnormalities of the foot. The mean age
of the donors was 65 years (range, 62 to 72 years). The
specimens were stored at -20°C and thawed at room
temperature 24 hours before the experiment. Each
specimen was disarticulated at the knee joint, and the
skin, subcutaneous tissues, and muscles were removed
from around the tibial and fibular shafts, exposing the
remaining tendons down to the level of the hindfoot.
Because the tibialis anterior, extensor digitorum longus,
and extensor hallucis longus tendons are inactive in the
midstance portion of the gait cycle, they were cut at the
insertion sites. The Achilles, peroneus longus, peroneus
brevis, flexor digitorum longus (FDL), and flexor hallucis
longus (FHL) tendons were sutured for reinforcement
and grasped by specially designed tendon clamps. The
proximal tibia and fibula were then firmly mounted in the
specimen-holding jig, which had several holes on the
top through which to pull the tendons. For the approxi-
mation of foot position at the midstance phase, an align-
ment apparatus was used to ensure consistent vertical
orientation of the tibia and fibula during fixation (Fig. 1).
Fig. 1. Experimental set-up showing a specimen with tendon
cables attached.
EXPERIMENTAL FLATFOOT MODEL 221
Calculation of Tendon Strength during Midstance
For each specimen, the tibialis anterior muscle was
stripped from its proximal to distal ends. This muscle
was chosen because it is superficial and easily sepa-
rated from the surrounding extensor muscle and fascia.
An electronic balance (Satorius AG Gottinqen", type B
410, Germany) measured the weight of the muscle. The
weight was divided by the specific density of muscle
(1.02 qm/cm') to calculate the volume, and then a
cross-sectional area was computed by dividing the vol-
ume by the mean fiber length of the tibialis anterior
muscle. According to the report by Steindler," the force,
or tension-producing capability, of the physiologic
cross-sectional area of the muscle is 3.6 kq/crn" of the
cross-sectional area. The maximal muscle force of the
tibialis anterior muscle is the cross-sectional area times
3.6 kq/crn", The maximal muscle force of the Achilles,
peroneus longus and brevis, FDL, and FHL tendons
was calculated according to the relative strength per-
centage from the study by Silver et al." Based on that
study, during the midstance phase of the gait cycle the
Achilles tendon is activated by 80%, the peroneus
longus tendon is activated by 40%, and the peroneus
brevis, FHL, and FDL tendons are each activated by
50% of the maximal strength of each tendon.' The force
applied on the different tendons was therefore the max-
imal muscle force adapted for its percentage activation
during the midstance phase of the gait cycle.
Method of Loading and Tendon Pulling
The sample was attached to a Mini Bionix load frame
(MTS Systems Corporation, Eden Prairie, MN) and, at
the upper cross-head, individual pneumatic cylinders
were attached. Wire cables were used to connect the
tendon clamps to the pneumatic cylinders. The tendons
were activated for the dynamic model and inactivated
for the static model. Based on the study by Silver et al.,"
the simulated muscle force applied to each tendon was
40% of the maximum muscle force, as calculated
above: FDL, 101 N; FHL, 101 N; peroneus brevis, 101
N; peroneus longus, 81 N; and soleus, 162 N. Muscle
force was limited due to the capacity of the tendon
clamps. The axial compression load selected was 280
N (40% of body weight). To simulate the clinical effects
of flatfoot and to observe changes in the arch over time,
cyclic loading with axial compression was applied
according to the experimental protocol below.
Stages of the Experiment and Fatigue Axial Loading
The experiment was divided into four stages:
1) The 280-N axial load was applied to the foot with no
previous cyclic loading;
2) 3,000 cyclic axial compressions (mean, 735 N;
range, 70 to 1,400 N [two times body weight]; fre-
222 CHU, MYERSON, NYSKA AND PARKS Foot & Ankle InternationalNol. 22, No. 3/March 2001

quency, 5 Hz) were applied, followed by the 280-N load; 2a

3) The spring ligament, plantar fascia, and talonavicular
capsule were released, and then the 280-N axial load
was applied; and
4) Another 3,000 cyclic axial compressions were
applied, followed by the 28-N axial load.

At each stage, arch measurements with and without

loading of the five tendons (Achilles, peroneus longus,
peroneus brevis, FDL, and FHL) were recorded.
Anteroposterior and lateral radiographs of the foot were
obtained for each stage and condition, resulting in eight
sets of measurements for each foot.

Radiographic Measurements
Talar-first metatarsal angle. This was measured on a
lateral radiographic view using the midpoint of a per-
pendicular line drawn from the dome of the talus to the
inferior beak of the lateral process of the talus. A second
point, which bisected the neck of the talus, was marked,
and a line connecting these two points was used as the
axis of the talus (Fig. 2a). A perpendicular line to the
articular surface of the base of the first metatarsal base
served as the longitudinal axis of the first metatarsal.
The angle between these two lines was defined as the
talar-first metatarsal angle.
Talocalcaneal angle. This was measured on a lateral
view. The midpoint between the inferior end of the medi-
al process of the calcaneal tuberosity and the superior
posterior corner of the calcaneus was marked. The mid-
point between the posterior edge of the posterior facet
and the shortest distance to the base of the calcaneus
was marked. The line connecting these two points served
as the axis of the calcaneus, and the angle between this
line and the talar axis was measured (Fig. 2b).
Talonavicular coverage angle. This angle was meas-
ured on an anteroposterior view. The angle was deter-
mined as follows. One line was drawn from the medial
and lateral margins of the articular surface of the navic-
ular, and a second line was drawn across the head of
the talus. A perpendicular line is drawn from each of
these two lines to the center of the navicular and the
center of the talar head, respectively. The intersection of
these two perpendicular lines creates the talonavicular
coverage angle, which increases as peritalar subluxa- Fig. 2. Artist's sketches of measurements made. a, talar-first
tion of the talonavicular joint worsens (Fig. 2C).12 metatarsal angle; b, talocalcaneal angle; c, talonavicular coverage
angle; d, cuneiform height.
Height of medial cuneiform. This was measured on a
lateral view. The shortest distance of the inferior surface
of the medial cuneiform from the line connecting the
inferior end of the medial sesamoid and the inferior determine if significant differences (P < 0.05) existed
point of the calcaneal tuberosity (Fig. 2d). between the groups. If a significant difference was
observed, least significant difference multiple range
Statistical Analysis tests were conducted to determine between which
A one-way analysis of variance (ANOVA) was used to groups the differences occurred.
Foot & Ankle InternationalNol. 22, No. 3/March 2001 EXPERIMENTAL FLATFOOT MODEL 223
10,------------- ---. 30,-------------------,

~
~ 20
~
"0
<Ii
g>
eo
L-
eo
B
.~
10

c:
o
~
c:
eo
Q)
::E 0
S1S S1D S2S S2D S3S S3D S4S S4D

Stage Stage

Fig. 3. Mean talar-first metatarsal angles at each stage and condi- Fig. 4. Mean talonavicular angles at each stage and condition test-
tion tested. 515, stage 1, static. 510, stage 1, dynamic. 525, stage ed. 515, stage 1, static. 510, stage 1, dynamic. 525, stage 2, stat-
2, static. 520, stage 2, dynamic. 535, stage 3, static. 530, stage 3, ic. 520, stage 2, dynamic. 53S, stage 3, static. 530, stage 3,
dynamic. 545, stage 4, static. 540, stage 4, dynamic. dynamic. 545, stage 4, static. 540, stage 4, dynamic.

RESULTS Height of the Medial Cuneiform

The results are presented graphically in Figures 3
through 6. In general, significance was observed only in
stage 4 results, after the supporting structures were cut
and after cyclic loading. We did not observe any signif-
icant difference between stage 4 static and dynamic
conditions, i.e., with or without tendon forces applied
about the ankle joint.
Talar-First Metatarsal Angle
There was a slight decrease, 1° (-9 to 10°) to -1° (-16
to 11 0), in the angle when applying the dynamic force
and after cyclic loading (stage 2); however, it did not
reach statistical significance. Interestingly, cutting the
spring ligament and the plantar fascia, applying axial
load, or applying axial load and dynamically loading the
tendons also did not significantly change this angle: 5°
(-10 to 16°) to 0.14° (-10 to 12°). However, after apply-
ing cyclic loading (stage 4), the angle decreased signif-
icantly from 5° (-10 to 16°) to -17° (-40 to 6°) for the stat-
ic condition, and from 2.5° (-11 to 15°) to -25° (-47 to
-11°) for the dynamic condition (Fig. 3).
Talonavicular Coverage Angle
Again, there were no significant differences observed
during the first three stages of the study, even after cut-
ting the spring ligament and plantar fascia. The angle
increased from 4.5° (0 to 20°) for Stage 1 to 6.5° (1 to
20°) for Stage 2 and increased to 10.5° (1 to 31°) for
Stage 3. At Stage 4, the angle increased to 22.8° for the
static condition. Again, after cyclic loading, significant
differences were observed (Fig. 4).
There was a gradual, slight decrease (without statisti-
cal difference) in the first three stages of the experiment
and when applying dynamic forces. When compared
with stages 1, 2, and 3, the height of the cuneiform after
cyclic loading (stage 4) decreased significantly in the
static condition and decreased even more in the
dynamic condition (Fig. 5).
Talocalcaneal Angle
There was a slight decrease in this angle in the
dynamic condition of stage 4, but it did not reach statis-
tical significance. There were no statistical differences
in this angle in the different stages or the different con-
ditions (Fig. 6).
DISCUSSION
In the current study, axial loading of the foot in either
the static or dynamic condition did not significantly
change the height of the arch, nor did the addition of
3,000 load cycles. These findings contrast with those of
Kitaoka et al.," who reported visible (but not quantified)
deformation of the arch during axial loading. However,
such small but significant changes in the height of the
arch are clearly a function of more precise evaluation of
the foot via a sonic digitizer or magnetic tracking sys-
tem. 6,7,15
From a clinical standpoint, evaluation of the arch in
patients is performed using radiographic measure-
ments, including the talonavicular coverage angle
(measured on anteroposterior radiographs), talocal-
caneal angle, medial cuneiform height, and talar-first
224 CHU, MYERSON, NYSKA AND PARKS
2.6,---------------------,
2.4
2.2
E 2.0
E g>
:i
.~ 1.8
.J::.
E ~
g 1.6
(1)
c:
:::l
o 1.4
c:
III
(1)
:a: 1.2
S1S SlO S2S S2D S3S S3D S4S S4D
Stage
Fig. 5. Mean cuneiform height measurements at each stage and
condition tested. 818, stage 1, static. 810, stage 1, dynamic. 828,
stage 2, static. 820, stage 2, dynamic. 838, stage 3, static. 830,
stage 3, dynamic. 848, stage 4, static. 840, stage 4, dynamic.
metatarsal angle (the latter three measured on lateral
radiographs). Despite the fact that more sophisticated
means of measurement were not available to us, it was
with this in mind that we adopted a more simplistic and
inexpensive model, in the hope that a comparison of
static and dynamic loads could be reproduced easily in
other laboratories that do not use magnetic tracking or
digitizers. Obviously, only substantial changes in posi-
tion can be evaluated by our system. The most signifi-
cant changes were observed in the talar-first metatarsal
angle and the talonavicular coverage angle, which were
indicated as sufficient for the evaluation of flatfoot defor-
mity. Although we observed a trend toward decreases in
the talocalcaneal angle, the minimal changes and large
standard deviation indicate the inaccuracy of this meas-
urement. In patients with long-standing deformities, the
talocalcaneal ligaments (mainly the interosseus)
stretch, increasing the valgus deformity. To create their
flatfoot model, Deland et al.' cut this ligament. The
insignificant change of the talocalcaneal angle in our
experiment indicates this ligament to be stable and
functioning even after cyclic loading.
We created our flatfoot model by sectioning the medi-
al supporting ligaments. The spring ligament, the long
and short plantar ligaments, and the plantar fascia are
the main static stabilizers of the arch. The major dynam-
ic contribution is by the tibialis posterior tendon, but the
peroneus longus and FDL tendons are secondary con-
tributors." In our model, although the arch tended to
lower after the static stabilizers were sectioned, the dif-
ference was not statistically significant. We think that
the varying elasticity, joint laxity, and bone anthropome-
Foot & Ankle InternationalNol. 22, No. 3/March 2001
134,----------------------,
133
~
~
~ 132
-c
<Ii
131
III
~
c: 130
§
~ 129
c:
~
::E 128
S1S S1D S2S S2D S3S S3D S4S S4D
Stage
Fig. 6. Mean talocalcaneal angles at each stage and condition test-
ed. 818, stage 1, static. 810, stage 1, dynamic. 828, stage 2, stat-
ic. 820, stage 2, dynamic. 838, stage 3, static. 830, stage 3,
dynamic. 848, stage 4, static. 840, stage 4, dynamic.
try in the cadaver samples contributed to the large stan-
dard deviation. Only after cyclic loading could we
demonstrate significant flattening of the arch, as indi-
cated by the talar-first metatarsal angle, suggesting that
this is the most sensitive measurement. Cyclic loading
contributed to additional stretching of the remaining lig-
aments.
In a clinical situation, when there is a rupture of the
tibialis posterior tendon, gradual flattening of the arch
takes place; but tearing of the static stabilizers of the
arch and inactivation of the tibialis posterior tendon may
eventually result in flatfoot.
In the current study, we demonstrated flattening of the
arch after 3,000 loading cycles when medial structures
were sectioned, In the study by McCormack et al.," flat-
foot was not produced until 12,000 cycles had been
applied, but those investigators sectioned only the
spring ligament. The additional cycles required were
probably due to stretching of the remaining stabilization.
Because the process of flattening the arch in patients is
usually gradual, it can be assumed that when static and
dynamic stabilizers of the arch are not functioning, fur-
ther full-weightbearing walking without support can
cause additional deterioration, as was shown by our
model. To differentiate between the importance of each
mechanism, the dynamic or static stabilizers of the arch
need further investigation.
In a cadaver model of the normal foot, Thordarson et
al." found that the Achilles tendon had a deforming
effect on the arch and that the FDL, FHL, and peroneus
longus tendons had small arch-supporting effects, In a
similar model, Kitaoka et al.' found that arch instability
Foot & Ankle InternationalNol. 22, No. 3/March 2001
was accentuated by loading the extrinsic tendons of the
foot. In our model, activation of the Achilles, peroneus
longus, peroneus brevis, FHL, and FDL tendons in the
dynamic phase increased the deformity of the flatfoot
after all the medial stabilizers were severed. Significant
flattening of the arch (by 57%), as measured by medial
cuneiform height, occurred in the dynamic condition
after cyclic axial loading. Activation of the tendons
before sectioning the medial stabilizers did not change
the arch height. Therefore, it appears that the Achilles
tendon plays a major role in the forces that deform the
arch, and that the combination of long flexors and per-
oneals cannot overcome its activity. The Achilles
deforming force is more pronounced when the static
stabilizers and tibialis posterior are torn and unable to
resist the deforming force. It may be that, in patients
with flatfoot, the unopposed force of the Achilles tendon
causes further deterioration, leading to its contracture.
In such patients, any physiotherapy that activates the
Achilles tendon may increase the deformity.
In conclusion, we have shown that, to create flatten-
ing of the plantar arch, there is a need to cut the medi-
al structures, includinq the spring and plantar ligaments
and possibly the plantar fascia. Cyclic axial loading after
destabilizing the foot increases the flattening of the
arch. Our results show that additional deterioration of
the arch is caused by dynamic loading of the flatfoot,
which may be attributed to the Achilles tendon. We did
not test deformity with or without Achilles load, but
always included other tendons. We believe that the
Achilles is a deforming force, although this conclusion
cannot be drawn from our data. In our model, we have
shown that activation of all the flexor muscles, including
the FDL, FHL, and peroneals, together with the Achilles
tendon, cause additional deterioration of the arch.
REFERENCES
1. Borton, D. C., and Saxby, T. S.: Tear of the plantar calcaneonav-
EXPERIMENTAL FLATFOOT MODEL 225
icular (spring) ligament causing flatfoot. A case report. J. Bone
Joint Surg., 79B:641-643, 1997.
2. Crowninshield, R. D., Brand, R. A., Johnston, R. C., and
Pedersen, D. R.: An analysis of collar function and the use of tita-
nium in femoral prostheses. Clin. Orthop., 158:270-277, 1981.
3. Deland, J. T., Arnoczky, S. P., and Thompson, F. M.: Adult
acquired flatfoot deformity at the talonavicular joint: reconstruction
of the spring ligament in an in vitro model. Foot Ankle, 13:327-332,
1992.
4. Funk, D. A., Cass, J. R., and Johnson, K. A.: Acquired adult flat
foot secondary to posterior tibial-tendon pathology. J. Bone Joint
Surg., 68A:95-102, 1986.
5. Gazdag, A. R., and Cracchiolo, A., III: Rupture of the posterior
tibial tendon. Evaluation of injury of the spring ligament and clini-
cal assessment of tendon transfer and ligament repair. J. Bone
Joint Surg., 79A:675-681, 1997.
6. Huang, C. K., Kitaoka, H. B., An, K. N., and Chao, E. Y.:
Biomechanical evaluation of longitudinal arch stability. Foot Ankle,
14:353-357, 1993.
7. Kitaoka, H. B., Luo, Z. P., and An, K.-N.: Effect of the posterior
tibial tendon on the arch of the foot during simulated weightbear-
ing: biomechanical analysis. Foot Ankle Int., 18:43-46, 1997.
8. Kitaoka, H. B., and Patzer, G. L.: Subtalar arthrodesis for poste-
rior tibial tendon dysfunction and pes planus. Clin. Orthop.,
345:187-194,1997.
9. Mann, R. A., and Thompson, F. M.: Rupture of the posterior tib-
ial tendon causing flat foot. Surgical treatment. J. Bone Joint
Surg., 67A:556-561 , 1985.
10. McCormack, A. P., Niki, H., Kiser, P., Tencer, A. F., and
Sangeorzan, B. J.: Two reconstructive techniques for flatfoot
deformity comparing contact characteristics of the hindfoot joints.
Foot Ankle Int., 19:452-461, 1998.
11. Myerson, M. S.: Adult acquired flatfoot deformity. Treatment of
dysfunction of the posterior tibial tendon. J. Bone Joint Surg.,
78A:780-792, 1996.
12.Sangeorzan, B. J., Wagner, U. A., Harrington, R. M., and
Tencer, A. F.: Contact characteristics of the subtalar joint: the
effect of talar neck misalignment. J. Orthop. Res., 10:544-551,
1992.
13.Silver, R. L., de la Garza, J., and Rang, M.: The myth of muscle
balance. A study of relative strengths and excursions of normal
muscles about the foot and ankle. J. Bone Joint Surg., 67B:432-
437,1985.
14.Steindler, A.: Post-graduate Lectures on Orthopedic Diagnosis
and Indications, Springfield (IL), Charles C Thomas, 1950,
15.Thordarson, D. B., Schmotzer, H., and Chon, J.: Reconstruction
with tenodesis in an adult flatfoot model. A biomechanical evalua-
tion of four methods. J. Bone Joint Surg., 77A:1557-1564, 1995.