iated with the authors of this page or responsible for its content.
HOJ 200117
79
INTRODUCTION
Femoral shaft fractures account for 21,000 hospitalizations
yearly and for 1.6% of all bony injuries in children less than
18 years of age. While most pediatric femoral shaft fractures
heal without significant complications, optimal treatment for
children greater than six years of age remains controversial.
Treatment options include traction followed by spica cast
immobilization, external fixation, flexible intramedullary (IM)
rods, interlocking IM nails and compression plate fixation.
Each of these methods has been associated with complications.
The ideal method of fracture fixation in children should suf-
ficiently stabilize the fracture to allow rapid mobilization and
joint motion while avoiding injury to the growth plate. While
interlocking IM nails have been quite successful in adults,
they have been associated with trochanteric growth arrest and
osteonecrosis of the femoral head in children. Intramedullary
fixation with flexible rods is rapidly gaining acceptance for
the treatment of pediatric femur fractures since it avoids
these risks. However these flexible rods provide little rotational
stability and cannot be used in unstable fracture patterns (i.e.
large butterfly fragment of greater than 50% of the width of
the bone or with segmental comminution). We have adapted
a segmented, interlocking, titanium nail developed for treating
humeral diaphysis fractures in adults (Synthes, Paoli, PA) to
be used treating femoral diaphysis fractures in children six
years or older (Fig. 1). The aim of this ex-vivo, biomechanical
study was to compare the structural compliance, failure load
and ability of this segmented, interlocking nail to limit fracture
gap motion to current methods of fixing femoral diaphysis
fractures in children using either flexible intramedullary rods
or double-stacked, monolateral external fixators. Since cadaver
femurs from children were unavailable in adequate numbers to
conduct a study of sufficient statistical power, synthetic femurs
representing the average size and shape of childrens femurs
aged six to 13 years were used and subjected to multiaxial and
multimodal loading conditions.
METHODS
Proportionately-sized synthetic femurs (Pacific Research,
Vashon, WA) that model the structural properties of human
femurs were obtained in 29, 38 and 42 centimeter lengths
representative of femurs from six, ten and 13 year old children
respectively. The size and shape of the six-year-old synthetic
femur was derived from a plaster cast of a cadaver femur
obtained from the Smithsonian Institute. We validated that the
size, shape and bending rigidities of the 38-cm. and 42-cm
length synthetic femurs replicated those of 10 and 13 year
old children, respectively using three-dimensional computed
tomography scans of femurs from patients previously scanned
at Childrens Hospital as part of unrelated study.
An oblique, mid-shaft fracture was simulated by a 45
o

osteotomy in nine femurs of each length using a handsaw
and miter. Three femurs of each size were allocated to each
of three fixation groups: (1) segmented, interlocking 7.5 mm
diameter titanium IM nail (Synthes, Paoli, PA) inserted laterally
through the greater trochanter apophysis (Fig. 2); (2) two
narrow diameter (3.5 4.5 mm), flexible stainless steel Enders
IM rods (Smith Nephew, NJ) inserted retrograde through the
medial and lateral distal metaphysis; (3) four pin (4.5mm
F
EMORAL
S
HAFT
F
RACTURES

IN
C
HILDREN
: C
OMPARATIVE

B
IOMECHANICAL
A
NALYSIS

OF

A
N
EW
F
LEXIBLE
I
NTERLOCKING

N
AIL

TO
C
URRENT
T
REATMENT
M
ETHODS
*W B
ARTHOLOMEW
, BS; *S K
WAK
, P
H
D; ***D R
ING
, MD; ***B S
NYDER
, MD
*O
RTHOPEDIC
B
IOMECHANICS
L
ABORATORY
, B
ETH
I
SRAEL
, D
EACONESS
M
EDICAL
C
ENTER
, B
OSTON
, MA, ** D
EPARTMENT

OF
O
RTHOPAEDIC

S
URGERY
, C
HILDREN S
H
OSPITAL
, B
OSTON
, MA, *** D
EPARTMENT

OF
O
RTHOPAEDIC
S
URGERY
, M
ASSACHUSETTS
G
ENERAL
H
OSPITAL
, B
OSTON
, MA
Fig. 1 AO segmented flexible intramedullary interlocking titanium humeral nail.
Fig. 2 The segmented, IM
humeral nail was inserted in
the flexible state antegrade
through an elliptically shaped
lateral portal created at the
greater trochanter. Proximal
and distal locking screws were
placed under fluoroscopic
guidance. A drive screw was
then advanced to tension an
internal wire thereby com-
pressing and interlocking the
segments to create a rigid
construct.
W Bartholomew, BS, Research Staff, Orthopaedics Biomechanics Laboratories, Beth
Israel Deaconess Medical Center, Boston, MA.
S Kwak, PhD, Research Staff, Orthopaedics Biomechanics Laboratories, Beth Israel
Deaconess Medical Center, Boston, MA.
D Ring, MD is an Instructor in Orthopaedic Surgery at Harvard Medical School.
B Snyder, MD, PhD is an Assistant Professor in Orthopaedic Surgery at Harvard
Medical School.
Address correspondence to:
Brian D. Snyder, MD, PhD
Department of Orthopaedic Surgery
The Childrens Hospital
300 Longwood Avenue
Boston, MA 02115 80
Shanz screw), double stacked, carbon-fiber rod, AO external
fixator mounted laterally (Synthes, Paoli, PA). Since connective
tissues were absent, fracture stability was provided solely by
the hardware. The distal end of each femur was potted in
methylmethacrylate and the proximal end was potted in a
low melting point metal that allowed non-destructive removal
of the femoral head from the potting material. Fracture gap
motion was measured using two video cameras and a set of
four infrared reflective markers attached to both sides of the
osteotomy to monitor the rigid body motion of each fragment.
The maximum fracture gap displacement was determined
from the difference in rigid body motion of the two femur
segments in three dimensions. Specially designed loading
frames mounted in a servohydraulic testing machine (Instron
1331, Canton MA) operated under load control allowed the
application of four-point bending moments in the frontal and
sagittal planes (Fig. 3), torsional moments along the femoral
diaphyseal axis (Fig. 4) and a multiaxial, multimodal loading
configuration that simulated single legged stance (Fig. 5).
Allowing the femurs to warp or slide out of plane
eliminated the application of extraneous forces and moments.
Structural compliance was derived from the slope of the
displacement-load curve for each mode of testing. This param-
eter is the inverse of stuffiness and quantifies the overall
motion or displacement of the stabilized femur construct in
response to an applied load. The specimens were then loaded
to failure in single legged stance. Failure was defined as the
load at which the displacement-load curve deviated 2% from
the linear portion of the curve. Compliance, maximum fracture
gap displacement and failure load were compared for each type
of fixation and for each femur size for every loading mode.
Two-way analysis of variance was performed to examine the
effect of fixation type and femur size on compliance, fracture
gap displacement and failure load.
RESULTS
C
OMPLIANCE
Both femur size and fixation type affected structural
compliance in bending, torsion and single legged stance.
The effect of fixation type was influenced by femur size.
Intramedullary devices were more compliant in the larger
femurs. We found that the external fixator was the least
compliant for all testing modes and all femur sizes. The
segmented IM nail was less compliant than the Enders rods
in torsion and in single legged stance, but in bending the
segmented nail was different from the Enders rod only for
the smallest femur.
F
RACTURE
G
AP
D
ISPLACEMENT
Both femur size and fixation type affected fracture gap
displacement for flexion / extension bending in the sagittal
plane (Fig. 6). There was more gap motion in larger femurs.
There was also more gap motion for Enders rods than for the
segmented nail or external fixator. The fracture gap displace-
ment in flexion/extension bending was not significantly differ-
ent between the segmented nail and external fixator.
Only the type of fixation used affected fracture gap
displacement for varus/valgus bending in the frontal plane (Fig.
7). There was more gap displacement for Enders rods than
for the segmented nail or external fixator. The external fixator
Fig. 3 (left) A specially designed loading frame mounted in a servohydraulic materi-
als testing machine allowed the application of four-point bending in either the frontal
or sagittal planes. A non-destructive physiologic moment of 6.35 Nm was applied
for five cycles at 0.5 Hz.
Fig. 4 (right) A force couple applied using a cable drive system mounted in a
servohydraulic material testing system imposed a torque of 4.14 Nm for five cycles
at 0.5 Hz along the femoral diaphysis. This cable drive system allowed the femur to
warp out