Effects of Foot Posture on Fifth Metatarsal Fracture Healing: A Finite Element Study

Abstract
The goal of this study was to evaluate the effects of maintaining different foot postures
during healing of proximal fth metatarsal fractures for each of 3 common fracture types. A
3-dimensional (3D) nite element model of a human foot was developed and 3 loading
situations were evaluated, including the following: (1) normal weightbearing, (2) standing
with the affected foot in dorsiexion at the ankle, and (3) standing with the affected foot in
eversion. Three different stages of the fracture-healing process were studied, including: stage
1, wherein the material interposed between the fractured edges was the initial connective
tissue; stage 2, wherein connective tissue had been replaced by soft callus; and stage 3,
wherein soft callus was replaced by mature bone. Thus, 30 3D nite element models were
analyzed that took into account fracture type, foot posture, and healing stage. Different foot
postures did not statistically signicantly affect the peak-developed strains on the fracture
site. When the fractured foot was everted or dorsiexed, it developed a slightly higher strain
within the fracture than when it was in the normal weightbearing position. In Jones fractures,
eversion of the foot caused further torsional strain and we believe that this position should be
avoided during foot immobilization during the treatment of fth metatarsal base fractures.
Tuberosity avulsion fractures and Jones fractures seem to be biomechanically stable fractures,
as compared with shaft fractures. Our understanding of the literature and experience indicate
that current clinical observations and standard therapeutic options are in accordance with the
results that we observed in this investigation, with the exception of Jones fractures.

Fractures of the fth metatarsal are the most common metatarsal fractures in children over 5
years of age and adults. Approximately 45% to 70% of all metatarsal fractures involve the
fth metatarsal, and the incidence of this injury is estimated at 1.8 per 1000 persons each
year. These fractures can be classied into 3 types, namely tuberosity avulsion fractures,
fractures at the metaphyseal-diaphyseal junction (Jones fractures), and shaft stress fractures.

Various studies have evaluated the etiology of each type of fth metatarsal fracture, as well
as the best therapeutic options. As far as tuberosity avulsion fractures are concerned, if the
injury involves large displaced intra-articular fragments, surgical management is required,
whereas conservative management is the common treatment modality of all nondisplaced or
minimally displaced fractures. Treatment options are weightbearing as tolerated in a stiff-
soled shoe, plaster cast, or elastic dressing. Although the prognosis of the conservative
treatment of avulsion fractures is good and nonunions are uncommon, recovery can take up to
6 months or longer. Fractures at the metaphyseal-diaphyseal junction, also called Jones
fractures, are named after Sir Robert Jones. These are often nondisplaced fractures and their
treatment is controversial. Various experts recommend either weightbearing or non-
weightbearing in a short leg cast as well as open reduction and internal xation with lag
screws. Shaft stress fractures are considered pathological fractures that occur mainly in
athletes, although they can also occur in patients with poor bone stock. These are commonly
associated with repetitive distraction forces. If the patient is a competitive athlete or if the
fracture is displaced (more than 3 mm) or angulated (more than 10o dorsal or plantar),
surgical intervention may be required. Otherwise, nonsurgical management is followed and
consists of non-weightbearing immobilization for up to 3 months. Elastic dressing, a posterior
splint, short leg cast, or hard plastic cast shoe have also been proposed therapies for this
injury, even though they pose problems because of their tendency for nonunion, and patients
who disregard the postoperative non-weightbearing protocol are prone to experience fracture-
healing problems.
Loading of the foot is the culmination of a complex interaction between bone, cartilage,
muscle, ligament, fascia, and the external environment. To quantify foot biomechanics, many
experimental techniques have been developed, such as pressure-sensitive lms, pressure-
sensing platforms, in-shoe pressure transducers, cadaveric experiments, in vivo force
measurements, and gait analysis. These techniques are commonly used to predict joint
kinetics and quantify plantar pressure distributions. They provide important information on
overall reaction forces but cannot directly assess the internal stress state of the foot and its
components. To overcome these intrinsic difculties and to supplement the experimental
techniques, computer-modeling approaches such as nite element analysis (FEA) are
commonly used by engineers and other investigators. FEA calculates the load distribution
and displacement, and internal stresses and strains, of each component part being tested.
Furthermore, it allows researchers to analyze the 3-dimensional (3D) motion of the bone and
the deformation of soft tissues under specic conditions. A number of 3D FEA foot models
have been developed to study the biomechanical effect of various normal and pathological
conditions. However, few biomechanical studies concerning the fractured fth metatarsal
have been published.

The objective of our study was to assess whether there is a foot posture most conducive to
bone healing after a fracture of the fth metatarsal. We developed biomechanical parameters
for strain, stress, and displacement for each of the 3 common fth metatarsal fracture types.
Then we tested for an association between the fractures biomechanics, foot position, and
clinical outcome through an analysis of 30 3D FEA foot models, taking into account the 3
different types of fractures, foot postures, and stages of bone healing. Our ultimate goal was
to attempt to shed light on the optimal treatment strategy for common fractures of the fth
metatarsal.

Materials and Methods

Finite Element Models
A 3D foot finite element (FE) model was developed based on a volume reconstruction of the
coronal computer tomography (CT) images of the right foot of a healthy male volunteer (34
years of age, height 172 cm, weight 82 kg). CT scan properties were 0.45-mm resolution and
0.75-mm slice spacing. The CT dataset was imported into MIMICS version 10 medical image
processing software (Materialise, Leuven, Belgium). The images were segmented and the
boundaries of skeleton and skin surface were obtained. The boundary surfaces of the skeletal
and skin components were then processed with the pre-processing tool of the FE analysis
software ANSYS version 12.0 (ANSYS, Inc., Canonsburg, PA) in order to form solid models
for each foot and ankle bone and the entire external foot surface.

The FE model consisted of 28 bony segments: the distal parts of the tibia and fibula, talus,
calcaneus, cuboid, navicular, 3 cuneiforms, 5 metatarsals, and 14 components of the
phalanges. Sesamoids were not simulated separately; instead they were considered fused with
the first metatarsal, making an enlarged first metatarsal head. Cartilage properties were given
to the volumes that corresponded with their anatomical site. The surface-to-surface contact
option in ANSYSwas used to simulate the frictionless contact relationship between articular
surfaces. Fifty-three plantar and dorsal ligaments, along with plantar fascia, were designed by
connecting the corresponding attachment points on the bones. Each ligament was simulated
by one or multiple link elements based on its individual shape and size. To simulate the
plantar fascia, 5 tension-only link elements were used. The origin and the insertion sites of all
ligaments were based on their anatomical location according to an anatomy atlas. Bones,
cartilages, and soft tissue were meshed with a total of 172,000 ten-noded tetrahedral solid
elements (Fig. 1A). Link elements that have tension-only functional capability were used to
simulate ligaments bearing the tensile load. All bony and ligamentous structures were
embedded in a homogenous soft tissue mass that represented all other foot structures and
defined the external surface of the foot (Fig. 1B). A horizontal plate was used to simulate
ground support. According to the literature, a frictional contact interaction was established
between the plantar surface of the foot and the ground, and a coefficient of friction of 0.6 was
used.

Modeling of the 3 common fracture types of the proximal fifth metatarsal was done by
selecting the corresponding elements contained within the fracture site (Fig. 2). To analyze
the healing process, a modulus of elasticity ranging from 2 to 7300 MPa was attributed to the
elements to simulate the healing process. Healing consists of sequential steps of tissue
differentiation, bone absorption at the fractured surfaces, and uniting of the fracture
fragments by callus formation. Finally, the fracture undergoes long-lasting internal
remodeling. During this process, increasing increments in the modulus of elasticity, until
reaching the one corresponding to the normal bone, when the fracture is healed, represents
the material interpolated into the fracture site. Three different stages of the fractures healing
process were modeled. In the first stage, the material interposed between the fractured edges
was the initial connective tissue; in the second stage, connective tissue had been replaced by
soft callus; and in the third stage, soft callous was replaced by mature bone. Thus, 30 3D
finite element models were analyzed that took into account fracture type, foot posture, and
healing stage. For each stage, appropriate mechanical properties (Table 1) were assigned to
the material of the fracture site. The time period between the injury and each one of these
stages was not taken into consideration because the investigation of the temporal evolution
was not an objective.

All tissues were defined as homogeneous, isotropic, and linearly elastic. The Youngs
modulus and Poissons ratio for the bony structures were selected by weighing cortical and
trabecular elasticity values. The mechanical properties of cartilage, ligaments, and plantar
fascia were also taken from the literature. The average elastic modulus for the soft tissue
mass was adopted from measurements of an ultrasonic indentation system. All reference
material properties and element types are listed in Table 2. A mesh sensitivity analysis was
performed to ensure that the mesh density used in the FE model was sufficient to reach the
converged numerical results and that no further mesh refinement was necessary.

Loading Conditions
The proximal portion of the model was fully fixed to simulate the effects of constraints from
proximal tissues. The following 3 loading situations were examined:
1. Normal weightbearing
2. Standing on a dorsiflexed foot (dorsiflexed at the ankle)
3. Standing on an everted foot

These foot positions represent the hypothetical scenarios in which a patient with a fifth
metatarsal fracture is standing upright bearing his weight while his foot is immobilized.
Normal stance represents the usual position of the foot when it is immobilized in a short leg
cast or in orthotics while bearing weight. The other 2 loading cases were selected as
alternative immobilizing positions of the foot. In addition, considering that no immobilization
(cast or orthotics) has been used, all 3 loading cases are positions of the foot during the gait
cycle.
For each one of the loading cases, the following force vectors were applied to the foot:
1. The ground reaction force that corresponds to half of the body weight (410 N)
2. The reaction of the Achilles tendon
3. The reaction of the peroneus brevis muscle

Ground reaction force was applied to the underside of the supporting plate. This force was
applied at the center of the pressure applied to the plantar surface of the foot as it was defined
from foot pressure analysis measurements. An exceptionwas made for the everted foot, in
which 50% of the net ground forcewas applied to the first submetatarsal region and the other
50% to the subcalcaneal region. The net ground force vector during normal weightbearing
was applied perpendicular to the ground, whereas for the other 2 cases this vector was applied
at an angle of 20_ in relation to the forefoot for the dorsiflexed foot and 10_ in relation to the
calcaneal axis for the everted foot. For all 3 loading situations, the point of the ground force
application was constrained to allow

movement in the vector direction only. A force of 205 N was applied, in accordance with a
study by Simkin, who reported that the Achilles tendon force is approximately 50% of the
force applied to the foot during balanced standing. This force was defined by 5 equivalent
force vectors and was applied at the posterior extreme of the calcaneus. The peroneus brevis
tendon inserts in a fanlike pattern across the proximal fifth metatarsal and has been
implicated in avulsion-type fractures. Based on other studies, a value of 13.16 N was adopted
for the peroneus brevis muscle force applied on the proximal portion of the fifth metatarsal. A
3D free-body diagram depicts the boundary and loading conditions for each one of the
different postures studied (Fig. 3).
A parametric analysis comparing the 27 different models developedwas carried out to
determine the effects of different loading conditions on the mechanical environment of the
proximal fifth metatarsal during the healing process. This analysis allowed us to analyze
multiple cases by varying different parameters with regard to loading conditions, fracture
type, and stage of healing. Models of a healthy, nontraumatized fifth metatarsal for each
loading condition (3 models) were also created. Thus, a total of 30 FE models were studied.
The maximum von Mises stress on the metatarsals and the maximum principal strain of the
fractured site of the fifth metatarsal were recorded. The clinical significance of the peak
principal strain that is developed during the healing process has been described by Frost.
According to their studies, bone remodeling is promoted when the strain developed varies
between 50 and 3000 m strain. Strain values outside this zone do not promote fracture
healing. Thus, the maximum von Mises stress is used to estimate yield failure criteria in
ductile materials and calculate fatigue strength. The maximum stress developed inside a
structure should be below the maximum von Mises criterion to ensure that plastic
deformation does not occur.

Foot Plantar Pressure Measurement

In order to ensure the accuracy of our model, the FE predicted pressure distribution on the
plantar surface of the foot was compared with the corresponding value from the EPS Platform
(LorAn Engineering, Bologna, Italy) (Fig. 4). Foot pressure analysis was conducted on the
same 34-year-old healthy male volunteer (172 cm, 82 kg). Pressure between the ground and
the plantar surface of the foot was measured when the volunteer was standing upright and
barefoot on the platform. The plantar pressure distribution, peak plantar pressure, and center
of pressure were recorded with specific software (Footchecker version 4.0; LorAn
Engineering). Six clinically significant regions of interest were selected on the plantar surface
of the foot for comparison between the FE model predicted and the foot scan measured
results. These were the 5 metatarsal heads (#1 to #5) and calcaneal tuberosity (#6) (Fig. 5).

Results

Peak Principal Strain

For tuberosity avulsion fractures (Fig. 2A) with the individual standing upright and at the first
healing stage, when connective tissue has been interposed between the fractured bony parts,
the peak principal strain was 3060 m strain. At the second healing stage, when soft callus had
been formed, the value of the strain was 320 m strain, decreasing to 100 m strain at the third
healing phase of mature bone formation. When the same individual was standing on a
dorsiflexed foot, the peak principal strain was 3000 m strain, 330 m strain, and 100 m strain,
respectively. Finally, while standing with his foot everted, the respective values were 3170 m
strain, 350 m strain, and 100 m strain (Fig. 6).

For fractures at the metaphyseal-diaphyseal junction (Fig. 2B) and for each one of the 3
healing stages selected, the peak principal strain was 3120 m strain, 490 m strain, and 120 m
strain with the patient standing upright; 3180 m strain, 450 m strain, and 120 m strain with
the patient standing on a dorsiflexed foot; and 3160 m strain, 710 m strain, and 170 m strain
with the patient standing with his foot everted (Fig. 7).

Finally, for a shaft stress fracture (Fig. 2C), the peak principal strains for the healing stages
selected were 5460 m strain, 2400 m strain, and 750 m strain, with the patient standing
upright; 5620 m strain, 2450 m strain, and 770 m strain, with the patient standing with his
foot in dorsiflexion; and 5700 m strain, 2480 m strain, and 780 m strain with the patient
standing on an everted foot (Fig. 8).

Regardless of the type of fracture or the loading condition, the value of strain recorded
decreased as healing progressed. Throughout the fracture-healing process, the peak principal
strain, which was developed within the fracture site, was higher in shaft stress fractures than
in Jones fractures and tuberosity avulsion fractures. However, the percentage peak strain
reduction during fracture healing was different for each fracture type. This is the percentage
of difference of peak principal strain between the first and the second healing stage.

Specifically, percentage of peak strain reduction was 56% for shaft stress fractures, 85% for
Jones fractures, and 90% for tuberosity avulsion fractures. These values were similar for shaft
stress fractures and tuberosity avulsion fractures but not for Jones fractures. For this fracture
type, the reduction was lower when the foot was everted (74%). In addition, the respective
percentage of peak strain reduction between the first and the third (final) healing stage was
86% for shaft stress fractures, 96% for Jones fractures, and 97% for tuberosity avulsion
fractures.

Maximum von Mises Stress

The developed von Mises stress of the fifth metatarsal was slightly influenced by the different
loading situation or the different healing stage. In general, the maximum von Mises stress
developed in the fractured fifth metatarsalwas similar in Jones fractures and tuberosity
avulsion fractures. The bone with a shaft stress fracture developed lower von Mises
maximum stress compared with the other 2 fractures types, but this difference had no clinical
significance because the stresses developed within the bone were not so high as to affect its
stability (Fig. 9).

Discussion
Tuberosity Avulsion Fractures
Our study suggests that during the first healing stage, when the fibrous tissue is interposed
between the fractured bony parts, the value of strain developed does not promote bone
remodeling (Fig. 6), if the patient bears his weight regardless of the loading condition of his
foot (normal, dorsiflexed, or everted). When an individual is standing still, each foot supports
50% of the body weight. Moreover, if this patient walks in full weightbearing gait, the value
of strain developed is multiplied. However, in later healing stages, when soft callus has
formed, the strain developed within the fracture does not inhibit the healing process even if
the patient walks in a full weightbearing gait. Thus, according to our model, tuberosity
avulsion fractures with no displacement are biomechanically stable fractures. A small period
of non or partial weightbearing with crutches is needed for the initial management of these
fractures. The determination of the required time period would be very interesting, but our
model is not capable of estimating it.

Fractures at the Metaphyseal-Diaphyseal Junction (Jones Fractures)

Several authors have reported that Jones fractures are biomechanically unstable, because the
fractured area (meta-diaphyseal junction) is between the fixed base and the loose shaft. The
results of our study run counter to this, because in our model Jones fractures and tuberosity
avulsion fractures had no differences in biomechanical behavior. In fact, the peak stress
developed in the fifth metatarsal with a Jones fracture was similar to the peak stress that
developed in the fifth metatarsal with an avulsion fracture. Moreover, the graphs that depict
the strain developed within the fractured bony parts (Fig. 7) are similar to those of the
tuberosity avulsion fractures (Fig. 6), indicating that the mechanical behavior of these 2
fracture types is similar. Thus, the aforementioned analysis of the tuberosity avulsion fracture
management would be expected to apply to Jones fractures too. However, more authors
recommend not only cast immobilization without weightbearing, but also prolonged recovery
time. This disagreement is due to the anatomical location of fracture healing. The
metadiaphyseal region is the area where the vascular supply enters the bone. An injury to this
area accompanied by vascular disruption results in limited blood supply and places these
fractures at risk for nonunion.

For the Jones fracture type, the reduction of the strain developed between the early healing
stage and the soft callus stage seems to differ in relation to the other fracture types, when the
foot is everted. This means that the action of peroneus brevis and the lateral band of plantar
aponeurosis in the fifth metatarsal cause further torsional strain of the fracture. Attention
should be paid so as not to immobilize the foot in eversion.

Shaft Stress Fractures

Our results indicate that there are biomechanical differences between shaft stress fractures
and other types of fractures of the fifth metatarsal. Throughout the fracture-healing process,
the peak principal strain developed within the fracture site (Fig. 8) was higher in shaft stress
fractures than in the other 2 fracture types (Figs. 6 and 7). Moreover, the peak principal strain
developed during the first healing stage was 5460 to 5700 m strain, much greater than the
value range (50 to 3000 m strain) that promotes bone remodeling. Thus, if an individual is
standing in full weightbearing, regardless of his foot posture, fracture healing is not
promoted. Moreover, if this patient walks in full weightbearing gait, the value of strain
developed is almost duplicated. When soft callus has been formed, the strain developed
within the fracture does not inhibit the healing process if the patient is standing in full
weightbearing, but does inhibit it if the patient walks in partial or full weightbearing gait.
Only in the third stage of healing, when mature bone has been formed, was the strain value in
the boneremodeling zone even if the individual was walking in full weightbearing. These
experimental data, which are in accordance with the clinical observations, confirm that shaft
fractures are biomechanically unstable, have a problematic healing process, and need
protection until a stiff callus is formed. Therefore, a long period of non-weightbearing
immobilization is needed for the management of these fractures.