For several decades, intramedullary nailing has been one of the most
frequently used methods for osteosynthesis of metaphyseal and diaphyseal
fractures of the
tibia1,2.
In Europe, unreamed tibial nailing is a standard treatment in trauma
surgery1-3.
In the United States, reamed tibial nailing is preferred; it provides a higher
stiffness of the bone-implant construct as a result of larger possible
diameters and larger surface-area contact between the bone and the
nail4. Compared with
reamed tibial nailing, unreamed tibial nailing spares the endosteum and its
share of the cortical blood
supply3,5-8.
The intramedullary pressure and heat generated by the insertion of a tibial
nail without reaming are significantly smaller (p = 0.01) than those resulting
from reamed tibial
nailing5,9.
These attributes of unreamed tibial nailing add up to superior cortical blood
circulation, a lower complication rate in terms of fat embolism and infection,
and overall favorable conditions for biological
healing5,8,10.
However, the limits of intramedullary nailing are reached in the treatment
of high-energy trauma with severe damage of the tibia and/or the surrounding
soft tissues and in the treatment of very proximal or distal
fractures11,12.
In a retrospective analysis of ninety-four patients treated with unreamed
tibial nailing, Goldhahn et al. found seventeen cases (18%) of delayed
fracture-healing2.
Consequently, the biological advantages of unreamed tibial nailing are
occasionally associated with specific disadvantages. To spare the endosteum,
there is direct contact between the bone and the nail in only two small areas:
at the isthmus of the biconcave narrowing of the bone and at the angulated
proximal tip of the nail. The stiffness of the bone-implant construct is
almost exclusively defined by the contact and the load transfer through the
locking
bolts13.
With simple diaphyseal fractures (AO Type 42-A), there may be some
additional fracture site constraint due to direct contact between the
fragments. However, multifragmented or comminuted fractures (AO Types 42-B and
C) do not provide this opportunity, and the situation is further aggravated in
fractures associated with severe damage of the soft
tissues14.
The mechanical conditions at a fracture site, such as the interfragmentary
movements and gap size, have a major influence on
bone-healing15,16.
The initial phase of bone-healing is known to be especially sensitive to the
biomechanical boundary
conditions17. If
interfragmentary movements are not within a certain range, delayed union or
even nonunion may
result18. To what
degree interfragmentary movements after stabilization with unreamed tibial
nailing fall within the range known to be "optimal" for
bone-healing in humans remains unclear.
In a study of sheep, the interfragmentary movements during the initial
phase of healing were significantly higher at sites stabilized with unreamed
tibial nailing than they were at sites stabilized with a monolateral external
fixator19. The
locking bolts allowed a torsional rotation of up to 10° and an axial
compressive displacement of up to 3 mm at the osteotomy site in vivo.
Consequently, nine weeks after the osteotomy, biomechanical and histological
analysis showed that the healing in the group treated with unreamed tibial
nailing was significantly inferior to that in the group treated with the
external fixator. In vitro testing revealed significantly lower
bone-implant-construct stiffness after the unreamed tibial nailing compared
with that following the external
fixation19,20.
These findings suggested that, if the interfragmentary movements associated
with unreamed tibial nailing could be reduced in vivo, the biological
advantages of the procedure could be combined with the advantages of an
"optimal" mechanical situation at the fracture site.
The hypothesis of the present study was that a reduction of
interfragmentary movements, especially of torsional rotation and bending
angles, allowed by unreamed tibial nailing would support the healing process
and thus lead to an improved healing outcome at nine weeks. The objective of
this study was to investigate healing over a period of nine weeks following
stabilization of an unstable tibial osteotomy site with unreamed nailing with
a modified tibial device that had angle stable holes for the locking bolts
(angle stable tibial nail) and to compare this healing with that after
stabilization with standard unreamed tibial nailing.
Materials
A commercially available 9-mm-diameter titanium tibial nail designed
to be inserted without reaming (Synthes, Bochum, Germany) was shortened to fit
the length of an ovine tibia. The nail has two proximal and two distal locking
holes in the mediolateral direction, with a distance of 20.0 mm between the
centers of the proximal locking holes and 30.0 mm between the centers of the
distal locking holes. A tibial nail of the same design and material, also
intended to be inserted without reaming, was modified with smaller, threaded
locking holes (Stratec Medical, Oberdorf, Switzerland)
(Fig. 1). The threads of the
locking holes corresponded to the threads of commercially available 3.9-mm
titanium locking bolts, so that the bolts fit exactly into the holes. This
nail was called the angle stable tibial nail. The titanium locking bolts were
used in all of the animals. The length of the bolts was determined
intraoperatively. A custom-made aiming device (Stratec Medical), slightly
modified compared with the standard insertion instrument used for the standard
nail, was constructed for the angle stable tibial nail, allowing exact
placement of both the proximal and the distal locking bolts. The device was
connectable to the standard insertion instruments for the standard nail.
Animals
Twelve female Merino-mix sheep (two to three years old) were divided into
two treatment groups. Preoperative radiographs of the right tibia were made in
both groups to ensure that the bone had an inner isthmus diameter of >9.4
mm and was of sufficient length. Six animals were treated with standard
unreamed tibial nailing, and six animals were treated with the angle stable
tibial nail.
All animal experiments were carried out according to the policies and
principles established by the Animal Welfare Act, the National Institutes of
Health Guide for the Care and Use of Laboratory Animals, and the national
welfare guidelines. The design of the surgical procedures was critically
reviewed and observed by the local legal representative (Landesamt für
Arbeitsschutz, Gesundheitsschutz und technische Sicherheit, Berlin: G
00079/03).
Surgical Technique
In all sheep, the nailing was performed on the right tibia with the animal
under general anesthesia and lying on its right side. The insertion point for
the nails was on the anterior part of the tibial plateau, with an approach
medial to the patellar tendon. The capsule of the knee joint was not opened.
The diaphysis was osteotomized, with use of an oscillating saw, through a
medial skin incision at a defined distance from the medial malleolus. The
osteotomy site was distracted easily, and separation was maintained with a
3-mm spacer. The 3-mm gap ensured that no contact between the fragments would
occur, thus creating an "unstable" situation. The saw teeth had a
thickness of 0.7 mm, but at least 1 mm of bone was removed by the sawing
process. A standardized gap was achieved with distraction rather than by
creation of the gap by bone removal alone, which would have been less
reproducible because of the challenge of creating two exactly parallel
osteotomies.
After conventional insertion, all nails were locked with mediolateral
bolts, two proximal and two distal. To model a clinically relevant surgical
setting, the locking bolts for the standard tibial nails were inserted
proximally with a standard aiming device and distally under radiographic
control. The radiographically controlled method was not sufficiently accurate
for the distal interlocking of the angle stable tibial nails. Thus, the
custom-made aiming device was used to locate both the proximal and the distal
locking-bolt holes. Careful attention was paid to the restoration of the
physiological rotation of the tibia and to the preservation of the soft
tissues to exclude influences on the process of bone-healing and to establish
standardized experimental conditions.
To allow measurement of interfragmentary movements, two percutaneous 2.5-mm
Kirschner wires were inserted into the medial cortical bone, one proximal and
one distal to the osteotomy site. Direct contact with the nail was avoided.
The Kirschner wires served as mountings for the reflective marker frames used
for the optical measurements of the interfragmentary movements.
All wounds were closed in layers and were covered with spray dressing,
sterile compresses, and elastic conforming bandages. Finally, anteroposterior
and lateral radiographs were made immediately postoperatively.
Postoperative Care
The sheep received flunixin meglumine (Finadyne), 1.5 mL subcutaneously, as
an analgesic for three to nine days (median, seven days) postoperatively. The
wounds were inspected daily until they were completely healed. Inspection and
cleaning of the Kirschner wire tracks with 0.1% ethacridinlactatmonohydrat
(Rivanol) and skin disinfectant were performed daily. The Kirschner wires were
protected with compresses and tubular bandages. The healing course was
monitored with anteroposterior radiographs once weekly to look for implant
failure or a fracture that would require an animal to be killed early to keep
it from suffering.
Three-Dimensional Stiffness of the Bone-Implant Construct In
Vitro
Prior to the in vivo experiments, the in vitro stiffness of the
bone-implant construct was determined in six planes. An osteotomy was
performed and the site was stabilized with either standard unreamed tibial
nailing or an angle stable tibial nail in two groups of six cadaver sheep
tibiae each. These operations were performed in a manner similar to the
technique used in the in vivo operations, including the use of Kirschner wires
for the fixation of the marker frames for optical measurements. The proximal
and distal ends of the tibiae were embedded in acrylate (Beracryl; W. Troller,
Fulenbach, Switzerland), and the soft tissues were protected by moistened
elastic conforming bandages, similar to those used in the postmortem
biomechanical testing.
Biomechanical tests under six loading conditions were performed in a
nondestructive manner in a Zwick-1445 mechanical testing machine (Ulm,
Germany)21. A
preload of 400 N was used for axial compressive displacement; a 5-Nm torsional
moment and 25-N axial preloads, for torsional rotation; a 4-Nm torsional
moment and 40-N axial preloads, for anteroposterior and mediolateral bending;
and a 6-Nm bending moment preload (rectangular in the transverse plane)
combined with a 40-N axial preload for anteroposterior and mediolateral shear
displacement. Interfragmentary movements during the biomechanical tests were
measured optically with a motion capture system (PCReflex; Qualisys,
Gothenburg, Sweden) and were used to calculate the stiffness of the
bone-implant construct (in newtons per millimeter and newtons per degree as
appropriate).
Ground Reactions
All sheep were trained preoperatively to walk over a gangway with an
integrated pressure-sensitive platform (Emed-SF4; Novel, Munich, Germany).
Data were recorded only if the sheep placed at least one foot entirely on the
platform. All measurements were performed at a frequency of 50 Hz. The
ground-reaction parameters measured preoperatively were considered to be the
individual reference values. A complete gait cycle included the measurements
of the forefeet and hindfeet on both sides (single or in pairs per side). A
measurement session consisted of at least seven successful gait cycles for
each animal.
Postoperative measurements were made on the third day and once weekly over
a period of nine weeks. The maximum force, the contact area, and the contact
time were recorded as absolute values and are presented as a percentage of the
preoperative values.
Interfragmentary Movements
Simultaneous measurements of the interfragmentary movements during walking
were performed with an infrared optical system (PCReflex), at a frequency of
60 Hz, allowing, according to the Nyquist theorem, a maximum movement
frequency of 30 Hz to be
recorded22. During
the measurement session, reflective marker frames were attached to the
Kirschner wires, and three-dimensional movements between the frames were
recorded. The three-dimensional distances from the frames to the center of the
fracture gap were measured on the postoperative radiographs. These data were
used to calculate the interfragmentary movements in six planes, with use of
matrix algebra. The accuracy of the total system is ±0.1 mm for axial
compressive displacement and for anteroposterior and mediolateral shear
displacement, and it is ±0.1° for torsional anteroposterior
rotation and mediolateral bending
angles17.
Postmortem Procedure
All animals were killed at nine weeks. The osteotomized and contralateral
tibiae of each sheep were explanted, with preservation of the surrounding soft
tissues. All implants were removed with a standardized procedure. The initial
extraction torque (in newton-meters) of the locking bolts in the angle stable
tibial nails was measured with a torque wrench to exclude the possibility of
seizing of the bolts.
Biomechanical Testing of the Nine-Week Callus
The proximal and distal ends of the tibiae were embedded in acrylate
(Beracryl) at a defined distance between the embedding blocks. Each pair of
tibiae were tested wet in vitro until torsional failure (torsional rotation
about the long axis), at a rate of 10°/min, and with an axial preload of
20 N. The torsional stiffness (rotational stiffness about the long axis in
newton-meters per degree) and the maximum torsional moment (in newton-meters)
were recorded as absolute values and were calculated as a percentage of the
values on the intact side.
Anteroposterior and lateral radiographs were made immediately post mortem,
after implant removal, and after torsional testing. On the radiographs made
after implant removal, the bridged cortices were counted by two independent
examiners. During the entire period after the animals were killed, the tibiae,
including their soft-tissue coverage, were stored in 0.9% saline solution, and
the soft tissues were protected by moistened elastic conforming bandages
throughout the entire testing procedure.
Histomorphometric Analysis
After biomechanical testing, the surrounding soft tissues were removed and
all callus regions were sawed in the frontal plane. The anterior portion was
embedded in polymethylmethacrylate (Technovit 9100; Heraeus Kulzer, Wehrheim,
Germany), and 6-µm serial histological sections were cut at four
standardized compound depths with 50 µm between each section plane. These
sections were stained with safranin orange-von Kossa stain or safranin
orange-fast green stain for differentiation of callus tissues. Computerized
histomorphometric analysis was performed with an image analysis system (KS400;
Zeiss, Jena, Germany). The region of interest contained the gap plus twice the
width of the gap in the proximal and distal directions. The quality and
quantity of the callus tissue were examined with respect to bone, cartilage,
and fibrous tissue
formation23. Tissue
differentiation was analyzed at several locations within the callus
(endosteal, periosteal, lateral, and medial). All histological measurements
were made in each of the four sections and were averaged for each animal.
Statistical Analysis
Because of the low number of animals per group (six), the data are
presented as medians instead of means. Thus, interquartile ranges (25% to 75%
quartile) instead of standard deviations were used. The data were interpreted
with use of the Mann-Whitney U test. Possible correlations were examined with
use of the Spearman rho test and were verified with a test for multiple linear
regression and scatterplots (three independent
examiners)24. SPSS
11.0 for Windows software (SPSS, Chicago, Illinois) was used for the
statistical analysis. To determine statistical differences between the two
groups over the course of healing, a two-factorial analysis of variance for
repeated measurements was performed with SAS software (version 8.2; SAS
Institute, Cary, North Carolina). The level of significance was set at p <
0.05.
Animals
The sheep treated with the angle stable tibial nail weighed a mean
of 64.4 kg (range, 61.8 to 67.6 kg), and those treated with standard unreamed
tibial nailing weighed a mean of 81.8 kg (range, 53.0 to 93.0 kg). This
difference was not significant (p = 0.150).
Ninety-three trials performed to search for correlations between the
collected data and the animals' weight showed only one correlation that could
be confirmed by multiple linear regression (Spearman rho, p = 0.050): that
between the weight and the contact time (ground-reaction force) at day 21.
However, that correlation was not seen in the respective scatterplots.
Three-Dimensional Stiffness In Vitro
The biomechanical testing of the bone-implant-construct stiffness in the
cadaver bones revealed higher values, in several directions, in the group
treated with the angle stable tibial nail
(Table I). The stiffness of the
angle stable tibial nail was significantly higher, compared with that of the
conventional nail, for shear displacement in the anteroposterior direction (p
= 0.004), for shear displacement in the mediolateral direction (p = 0.030),
and for bending in the anteroposterior plane (p = 0.025). The angle stable
tibial nail also showed a tendency for higher stiffness for torsional rotation
(p = 0.055).
Ground Reactions
All animals in both groups unloaded the treated hindlimb initially, and
they showed similar values for maximum force, contact area, and contact time
(Fig. 2). However, whereas the
group treated with the angle stable tibial nail regained almost normal
weight-bearing and gait patterns of the treated hindlimb by the end of the
study, the group treated with standard unreamed tibial nailing did not. Over
the healing period, the group treated with the angle stable tibial nail showed
significantly higher maximum forces (p = 0.006), larger contact areas (p =
0.017), and longer contact times (p = 0.013), indicating significant
differences in the use of the treated limb. These differences increased
throughout the examination period.
Interfragmentary Movements
Over the examination period, in vivo gait analysis showed significantly
smaller interfragmentary movements in all directions in the group treated with
the angle stable tibial nail compared with the group treated with standard
unreamed tibial nailing (Table
II). The largest differences were seen in the early postoperative
period (Fig. 3). Throughout the
examination period, the anteroposterior bending angle and torsional rotation
decreased significantly in both groups, while axial compressive displacement
decreased significantly only in the group treated with the angle stable tibial
nail (Table II). Shear
displacement in the anteroposterior and mediolateral directions and the
bending angle in the mediolateral direction did not change significantly in
the group treated with the angle stable tibial nail. Shear displacement in the
anteroposterior and mediolateral directions decreased in the group treated
with standard unreamed tibial nailing (Fig.
3), but the decreases did not reach significance.
Radiographic Findings
At nine weeks, there was a tendency (p = 0.070) for more bridged cortices
per animal to be seen in the group treated with the angle stable tibial nail
(median, 3.3 per animal) than in the group treated with standard unreamed
tibial nailing (median, 2.0 per animal)
(Table I). It was remarkable
that, on the mediolateral radiographs (anterior and posterior cortices), a
total of ten cortices were seen to be bridged in the group treated with the
angle stable tibial nail, but only four were seen in the group treated with
standard unreamed tibial nailing. All sheep with an angle stable tibial nail
had at least one bridged cortex in this plane, whereas four treated with
unreamed tibial nailing had no bridged cortex. On the anteroposterior
radiographs (medial and lateral cortices), the group treated with the angle
stable tibial nail were seen to have, in total, only one more bridged cortex
(eight bridged cortices) than the other group (seven bridged cortices).
In four of the six sheep treated with the angle stable tibial nail, small
metal splinters were seen in the surroundings of at least one locking hole.
Abrasion of the metal was noticeable in one other sheep in that group. These
radiographic observations were confirmed by equivalent macroscopic
observations post mortem. No implant failure (bending or breaking of bolts or
nails) was seen in either group.
Initial Extraction Moments of Bolts
The initial extraction moment of the locking bolts ranged between 0 and
3.20 Nm, with a median of 0.72 Nm (Table
I). Some of the bolt heads (including those with an initial
extraction moment of 0 Nm) had to be freed of callus.
Postmortem Biomechanical Testing of Torsional Strength
The postmortem biomechanical testing showed the torsional stiffness
(rotational stiffness about the long axis) relative to the intact side in the
group treated with the angle stable tibial nail (91.6%) to be significantly
higher (p = 0.004) than that in the group treated with standard unreamed
tibial nailing (75.7%) and the maximum torsional moment to be significantly
larger as well (68.3% compared with 53.7%, p = 0.037)
(Table I).
Histomorphometric Analysis
The total width of the callus and the width of the mineralized bone in the
group treated with the angle stable tibial nail were significantly smaller
than those in the group treated with standard unreamed tibial nailing (p =
0.015 and 0.041, respectively) (Table
I). The total connective tissue area was also significantly
smaller (p = 0.004) and the cortical mineralized bone area was significantly
larger in the group treated with the angle stable tibialnail (p = 0.026). The
percentages of the total callus area consisting of connective tissue and
mineralized bone differed significantly (p = 0.006) between the two groups
(Fig. 4).
The objective of this study was to investigate whether a reduced
level of interfragmentary movement, especially in torsional rotation and
bending, would provide better healing nine weeks after unreamed tibial
nailing. The study showed that the significant reductions in interfragmentary
movements produced by the higher bone-implant-construct stiffness provided by
the angle stable tibial nail led to histological, radiographic, and
biomechanical evidence of superior bone-healing at nine weeks.
The axial compressive displacement in the group treated with the angle
stable tibial nail was within the range of 0.2 to 1.0 mm, which is recognized
as promoting bone-healing in
animals25. The
influence of shear displacement on bone-healing is the subject of
controversy25,26.
The increased shear stiffness and the correspondingly reduced shear
displacement in the group treated with the angle stable tibial nail were
associated with an increased prevalence of cortical bridging visible on the
radiographs made post mortem. However, this increase was not symmetrical:
there was a much greater increase in bridging of the anterior and posterior
cortices than in bridging of the medial and lateral cortices
(Table I). This finding
suggests a strong effect of the bending stiffness of the bone-implant
construct, since only this parameter showed an asymmetric difference between
the two treatment groups (Table
I). A bending movement applied across the fracture gap produces
mainly axial movements at the cortices (compressive at one and tensile at the
opposite one).
The measurements of the interfragmentary movements were made while the
animals were walking, without any running, hopping, or other sudden movements
of high impact. Nevertheless, sudden movements due to the friction between the
implants and the bone probably occurred. Rubin et
al.22 reported that
under flexed-limb conditions in humans, which are relevant to the sheep
hindlimb, mechanical signals with frequencies of >25 Hz showed reduced
transmissibility down the lower limb. Therefore, our chosen frequency of 60 Hz
should have been sufficient to reflect the displacements relevant to
bone-healing that occurred.
It is likely that the higher stiffness of the angle stable tibial nail
directly influences the weight-bearing of the treated limb. The unloading of
the limb by the sheep was contrary to their natural behavior. The fact that
they did not use the treated limb is a sign of considerable pain and
discomfort resulting from instability. The return to full weight-bearing in
the group treated with the angle stable tibial nail clearly indicates a
restoration of the functionality of the limb within the observed period of
nine weeks. Furthermore, functional weight-bearing is known to promote
bone-healing27. The
greater stiffness of the angle stable tibial nails seems to allow early
functional weight-bearing without the disadvantages of excessive
interfragmentary movements. There was a correlation between the animals'
weight and the contact time at day 21 (p = 0.050). However, the contact area,
maximum force, and interfragmentary movements at this time-point showed no
correlation with the animals' weight. In addition, not one of the other
ninety-two parameters tested showed a correlation with weight, suggesting that
the correlation with the contact time at day 21 was coincidental.
Fixation leading to rather large interfragmentary movements may cause
repetitive microruptures in the
callus28.
Bone-bridging may have been interrupted more frequently, leading to pronounced
callus formation, in the group treated with standard unreamed
nailing29,30.
The histomorphometric analysis showed a significantly smaller callus width and
mineralized bone width in the group treated with the angle stable tibial nail.
The significant differences in the morphology of the callus are consistent
with the superior bone-healing seen radiographically as better cortical
bridging in that group.
The significantly higher postmortem torsional stiffness in the group
treated with the angle stable tibial nail is consistent with the results of
the radiographic and histomorphometric analyses. The reduction of the
interfragmentary movements led to superior bone-healing at nine weeks. The
stiffness of the bone and thus its ability to absorb physiological forces and
moments was regained.
There was no evidence of seizing or galling between the bolts and the
nails, which suggests that there was no bending of the bolts in the angle
stable tibial nails by nine weeks, as bending would have inevitably resulted
in the seizing of the bolts. However, it is not possible to predict the
long-term endurance of the implant. The metal splinters that were observed
could be associated with surface scratches, providing stress concentrations
that might lead to implant failure. However, there was no macroscopic evidence
of this, and the metal splinters were found in conjunction with obvious damage
to the threads at the tips of the locking bolts. In any case, the production
of metal splinters would not be acceptable in an implant for use in human
patients; thus, the interlocking procedure must be refined.
The finding of less optimal healing following the standard unreamed tibial
nailing may not be directly applicable to the clinical situation. Unreamed
tibial nailing has been frequently reported to be successful, especially in
less complex
cases2,12.
In the present study, we examined a noncomminuted 3-mm gap defect, but the
open osteotomy added a moderate component of soft-tissue damage directly
surrounding the gap. In order to keep the model as reproducible as possible,
the bone was not comminuted. The approximately 2-mm distraction length was
<0.1% of the tibial length. No effort was needed to achieve the distraction
in any animal. Thus, it seems unlikely that the distraction had an influence
on the mechanical conditions, even though such an influence cannot be fully
excluded.
We believe that, in combination with the sheep's
"noncompliance" in terms of weight-bearing, the presented model
represents a reasonably challenging situation for standard unreamed nailing.
Even though, from a biological point of view, the model does not fully reflect
the most severe clinical situations that can be encountered, with respect to
implant loading the unreduced defect that we studied is a worst-case
scenario13. In this
demanding setting, a reduction of the interfragmentary movements at sites
treated with unreamed nailing seems to combine the advantages of favorable
biological conditions with the mechanical advantages of systems with a higher
biomechanical stiffness. In the present in vivo animal study, an earlier and
more effective restoration of the functionality of the limb was achieved by
the reduction of the interfragmentary movements. Even though the experimental
findings appear very promising, the general concept of "optimized"
interfragmentary movements by the use of angle stable bolts in unreamed
nailing needs to be confirmed in a prospective randomized clinical
investigation. ?