A biomechanical evaluation of different fixation strategies for posterolateral fragments in tibial

plateau fractures and introduction of the ‘magic screw’

abstract

Background: Posterior plate fixation is biomechanically the strongest fixation method for
posterolateral column fracture (PLCF) of the tibial plateau; however, there are inherent deficiencies
and risks of a posterior approach. Thus, the ‘magic screw’ was proposed to enhance fixation stability
of the lateral rafting plate used for PLCF. The purpose of this study was to reexamine and compare
the stability of different fixation methods for PLCF.

Methods: Synthetic tibiae models were used to simulate posterolateral split fractures. The fracture
models were randomly assigned into three groups: Group A, fixed with posterolateral buttress
plates; Group B, with lateral locking compression plates (LCP); and Group C fixed with lateral LCPs
and one ‘magic screw’. Gradually increased axial compressive loads were applied to each specimen.

Results: There was a mean subsidence hierarchy of the posterolateral fragment at different load
levels: Group A had the least subsidence, followed by Group C, and Group B had the most. There
were no significant differences in the mean loads at different displacements between Group A and
Group C. Group A had the highest axial stiffness. Additionally, there was a significant difference in
axial stiffness between Group B and Group C.

Conclusion: Biomechanical stability of the combined fixation of the posteriorly positioned lateral
rafting plate with the ‘magic screw’ was much closer to that of posterior plate fixation for split-type
PLCF. The necessity of posterior fixation through a posterior approach may be reduced for selected
patients.

Introduction

There have been times in the past when a posterolateral column fracture (PLCF) of the tibial plateau
was considered a relatively uncommon injury, accounting for seven to 10% of tibial plateau fractures
[1–3]. However, a recent morphological study has shown an incidence rate of PLCF in bicondylar
tibial plateau fractures as 44.32% [4]. Based on another investigation, except for Schatzker type III
fractures, PLCF can be observed in each type of Schatzker fracture, and the incidence of PLCF may
account for 54.3% of all posterior fractures [5]. These epidemiologic data indicate that PLCF is more
common than previously thought [6]. Thus, the treatment of PLCF might play a significant role in the
overall effectiveness of tibial plateau fracture management.
As a part of the articular surface of the tibial plateau, the posterolateral column contributes to load-
bearing when knee flexion =90°. Underestimation or failure to identify this fracture at first injury
inevitably leads to improper treatment and serious dysfunctional consequences [4,6,7]. Due to the
morphological features of the posterior aspect of the proximal tibia [8], and anatomical complexity
of the posterolateral corner of the knee [9], surgical strategies for PLCF are complicated. In the past
decade, many surgeons have attempted and explored various posterior approaches to manage and
implant posterior buttress plates for PLCF [2,3,7,9–28]. Although the results of these different
surgical approaches with posterior plating fixation have been satisfactory thus far, through in-depth
studies the risks and deficiencies of the posterior approach are being gradually realized
[29],especially iatrogenic injuries to normal structures.

It is unlikely that there is only one treatmentmethod for any kind of disease, and non-posterior
treatment isworth attempting for PLCF [30–40]. According to experience, PLCF can be reduced using
the anterolateral or lateral approaches, and fixed by a posteriorly positioned lateral plate; however,
the number of screws on the plate that can support and maintain posterolateral articular fragment is
limited (as shown in Figure 1). If the lateral plateau main depression area is located anteriorly, then
the lateral rafting plate should be implanted anteriorly.

Currently, PLCF fixation strength may be insufficient. Thus, the current center attempted to reinforce
the holding power and stability of PLCF by adding an extra screw outside the lateral rafting plate,
and named it the ‘magic screw’ [29].Although it has been confirmed that posterolateral buttress
plate fixation is biomechanically the strongest method for the split type of PLCF, and this result has
great guiding significance for treatment options [41], a previous study also showed defects and
limitations. Taking the recent clinical practices into account, it is believed that it is necessary to re
verify the biomechanical stability of the existing fixation methods for PLCF, and explore the
biomechanical property and application value of the ‘magic screw’, especially the synergistic effect
with the lateral plate.

Materials and methods

Group

Twenty-four left, synthetic, adult-sized tibiae models (Synbone, type 1350. Synbone AG, Malans,
Switzerland) were used for biomechanical testing. Each model was made of a rigid foam cortical shell
with cancellous material within the proximal and distal ends of the model. Synthetic tibiae were
used to minimize variability between specimens and to provide a consistent specimen size. The
models were obtained from one manufacturing batch to ensure the same material property and
geometry. All the models were randomly assigned into three groups, and fixed with three different
internal fixation patterns.

Each of the 24 specimens was randomly assigned to one of the following three fixation methods. A
posterolateral buttress plate was used in Group A: a 3.5-mm six-hole metaphysical locking plate
(Synthes GMBH, Zuchwil, Switzerland) used as the posterolateral buttress plate was contoured and
implanted obliquely from the proximal lateral aspect of the posterior tibia to the distal medial aspect
of the tibia (Figure 2A–D). A lateral, 3.5-mm, L-shaped proximal tibia locking compression plate (LCP)
was used in Group B: the transverse arm of the L-shaped plate was four holes long and 10 mm wide.
The LCP was placed as posteriorly as possible and through the superior fibular head space. The
posterior rim of the transverse arm of this anatomical plate was aligned with the posterior edge of
lateral plateau (Figure 1). One study found that the average articular surface area of lateral plateau
that remained unsupported was 40% by existing plate systems produced by different manufacturers
[33]. According to the measurement results, by Sohn et al. [42], of lateral anteroposterior distance
(LAPD) value and surface areas of the posterolateral fragment, and the matching chart of the
existing lateral plate with the fracture model, there is a limited number of rafting screws (one or at
most two screws) from the posteriorly positioned lateral plate that can hold the posterolateral
fragment or support the reduced articular surface (Figure 1).

In the current study, a lateral, 3.5-mm, six-hole LCP proximal tibial plate and one ‘magic screw’ was
\used in Group C: the

‘magic screw’ was obliquely implanted from anterior-inferior-medial to the posterior-superior-lateral

direction. The entrance point

of the ‘magic screw’ was determined prior to placement and consistent among the models. The
entrance point of the ‘magic screw’

was five to 10 cm distal to the medial plateau of the tibia, and located at the midpoint of the
anterior and posterior borders of the

medial aspect of the tibia. The ‘magic screw’ was oriented towards the posterior edge of the
posterolateral fragment, and avoided

the rafting screws placed through the lateral plate (Figure 2E–H). In addition, the ‘magic screw’ was
implanted subchondrally, and

the posterior cortexwas not punctured. Before fracturemodeling, this hardwarewas initially
implanted on the intact unosteotomized

synbones, based on the grouping. The implants were all removed after the fixation positions had
been verified by radiography.
To ensure material and design consistency, the implants used in all three groups were made by the
same manufacturer

(DePuy-Synthes, USA, Paoli, PA). Based on a pilot study that assessed variation in fixation strength
measurements, eight specimens

were assigned to each fixation group. With eight specimens in each group, it is possible to detect a
difference of 100 N in the mean

fixation strength, with a power of 80% and alpha error at 0.05.

Fracture models

Fracture models simulating posterolateral split fractures in the lateral and bicondylar tibial plateau
fractures were created for

biomechanical testing. Figure 3 demonstrates the modeling of a PLCF, based on data from the
morphological measurements by

Sohn et al. [42]. The posterior femoral condylar axis (PFCA) was used for the reference line with
which to measure the variable

parameters on the axial image. The parameters for model construction included the lateral major
articular fracture angle (LMFA),

posterior major articular fracture angle (PMFA), LAPD, posterior horizontal distance (PHD) in the
horizontal plane, and posterior

cortical height (PCH) and sagittal fracture angle (SFA) in the sagittal plane. The LMFA is the angle
between the anterior fracture

line of the posterolateral fragment and the PFCA. The PMFA is subtended by the posterior traversed
line and the line perpendic-

ular to the PFCA. The LAPD is the vertical distance from the lateral exit point of the PLCF to the
anterior cortex of the fibular head,

which is a perpendicular line to the PFCA. The PHD is the horizontal distance between the medial
cortex of the fibular head and

the posterior exit point of the posterolateral column fracture, which is a line parallel to the PFCA.
The PCH is the distance from the

point of the most posterior plateau to the posterior cortical breakage point. The SFA is measured by
subtending the joint line of
the posterolateral fragment with the coronal fracture line. Detailed numerical values of these
parameters are shown in Table 1.

The fracture models were created using a thin blade saw. After fracture modeling, all hardware was
finally implanted in situ

with a respective fixation pattern.

Test

Each potted proximal tibia was placed vertically in a material-testing machine (Instron 5569, Instron,
Norwood, MA, USA) for

primary stability and fixation failure testing. The loadwas applied to the horizontally positioned tibial
plateau through a customblock

(stainless steel, two centimeters long and one centimeter wide), which was used to distribute the
load only on the posterolateral

fragment. Bluehill 2 software (Bluehill 2, 2.17.649, Instron, USA) was used for load control and data
collection. The Optotrak Certus

(Serial No. C3-04950, NDI, Inc.,Waterloo, Ontario, Canada)motion analysis systemwas utilized to
track displacement of the postero-

lateral fragment in three dimensions. The Optotrak Certus system has an accuracy of up to 0.1 mm
and resolution of 0.01 mm. Two

markers were used: one was attached to the posterolateral fragment and the other to the posterior
aspect of the tibial shaft at the

same level (Figure 4A, C, and E). Displacement in the vertical axis between these two markers was
analyzed.

After mounting each specimen on the machine, gradually increased axial compressive loads were
applied to each specimen with

a load speed of one millimeter per minute. Axial displacement from the initial position to the
maximum load was continuously

recorded using NDI motion sensors. Load–displacement curves were generated for each, and axial
stiffness was calculated from the

linear portion of the load–displacement curve. Four load levelswith a peak force of up to 200N, 600
N, 800 N, and 1000 Nwere chosen
because the pretest had led to the assumption that no failure of fixation would occur within these
limits. The maximum peak force

was set at 1000 N. Subsidence was defined as the displacement of the posterolateral fragment
(Figure 4B, D and F). The specimens

were loaded to failure, which was defined as a fragment displacement measurement N3 mm. The
loads on three other subsidence

levels (six millimeters, nine millimeters, and 10 mm) were also collected and compared.

Statistical analysis

One-way analysis of variance (ANOVA) was conducted to determine whether loading points and final
failure differed among

these fixation patterns. A Fisher post hoc least significant difference criterion was used to correct for
multiple group comparisons.

A P-value of b0.05 was set as a significant value. All statistics were computed with the use of SPSS
version 19.0 software (SPSS,

Inc., Chicago, IL, USA).

Results

Throughout testing, loading was continued while there was no failure of the bone-implant
construction (plate loosening or

deformation, and screw loosening or breakage). Thus, there was no marked decrease of load versus
displacement curve.

Subsidence of the posterolateral split fragment of tibial plateau

Subsidence of the tibial plateau posterolateral split fragment under four level axial loads is
summarized in Table 2 and shown

in Figure 5. Mean subsidence of the posterolateral split fragment at different load levels (200 N, 600
N, 800 N, and 1000 N) were
significantly smaller for the posterior plate fixation construct (Group A) than the lateral plate only
(Group B) (P b 0.05). However,

the difference between the lateral plate and ‘magic screw’ fixation (Group C) and the lateral plate
only fixation (Group B) was

also statistically significant at different load levels (P b 0.05). There was a hierarchy of mean
subsidence of the posterolateral

fragment: Group A had the least subsidence, followed by Group C, and Group B had the most
subsidence. The subsidence tenden-

cies of Groups A and C were relatively close in the line chart (P N 0.05).

Loads to displacement N3 mm and axial stiffness

The loads to displacement N3 mm (load to failure) and axial stiffness are shown in Table 2. The loads
to subsidence at different

levels (three millimeters, six millimeters, nine millimeters, and 10 mm) for the three fixation
constructs are shown in Figure 6.

Group A resisted significantly greater loads at different subsidence levels compared with Group B
constructs (P b 0.05). Group C also resisted more significant loads at different subsidence levels than
Group B (P b 0.05). There was no significant difference in the

mean loads at threemillimeters, sixmillimeters, ninemillimeters, or 10 mm of displacement when

comparing Group A with Group C

(P N 0.05). However, Group A had the highest axial stiffness; it was significantly higher than the
other two fixation constructs

(P b 0.05). There was also a significant difference in the axial stiffness between Groups B and C (P b
0.05), as shown in Figure 7.

Discussion

The clinical treatment of complex tibial plateau fractures is still a challenge [43]. Currently, the
treatment of PLCF is a primary

issue with tibial plateau fractures. There have been numerous studies that have explored various
different approaches and fixation
methods for PLCF. In terms of surgical approach, there are three kinds of approaches that can be
applied, including: the

posteromedial [10–16], posterolateral [2,3,7,17–28], and anterolateral or lateral approaches [29–

40]. The posteromedial approach

for PLCF has been primarily used for complex three-column fractures [10,12,14,15] or fractures of
the whole posterior column

[13]. Because of the resistance of the posterior muscle and neurovascular bundle, the tibial plateau
posterolateral articular surface

could not be fully exposed and managed [11,44]. Despite the capacity for anatomical reduction, stiff
posterior plate fixation, and subchondral fixation for posterior lateral articular surface fractures,
there are still many flaws in various posterolateral approaches

[29]. Although posterior buttress plate fixation can be performed through posteromedial or
posterolateral approaches, the poste-

rior approaches have obvious limitations in the management of PLCF. Other approaches and fixation
methods should also be

considered, as they may offer alternative options.

At present, lateral rafting plate fixation is one of themost valuablemethods for PLCF treatment,
except for posterior buttress plate

fixation. Although it has been proven that biomechanical stability of lateral plating fixation for PLCF
is inferior to posterior buttress

plating fixation [41], a number of surgeons have successfully used lateral rafting plate fixation
tomanage PLCF through various lateral

or anterolateral approaches [29].Most of these studies were case series or case reports and
technique trick introduction.Meanwhile,

failure cases with PLCF malunion treated by lateral plate fixation are often encountered [6]. Thus, it
was still controversial whether a

lateral rafting plate could provide sufficient stability to the PLCF; in this construct the proximal
locking screws are parallel with the

coronal fracture line and therefore, two locking screws at most are available to gain adequate
purchase into the PLCF. Since that is

so, the ‘magic screw’ technique was proposed to enhance fixation stability of the lateral rafting plate
for PLCF on the basis of analysis
and practical experience. The ‘magic screw’ is similar to but different from the kickstand screw,
which is employed in the proximal

locking plates for peritrochanteric femoral fractures [45].The ‘magic screw’ is implanted outside the
lateral plate and thereby plays

a separate role. Previously, the fixation method of lateral rafting plate with ‘magic screw’ technique
has been clinically used to

treat selected patients with PLCF [29]. Therefore, a biomechanical investigation would be necessary.

A literature review revealed that little biomechanical research has been conducted regarding fixation
for PLCF [41,46]. Clinical

practice indicates that there is a need to improve and further verify the details of biomechanical
testing for PLCF fixation, includ-

ing the fracture model, implants, loading modes, etc. The latest radiographic data from a
morphological study were used to create

the split type fracture model of PLCF [42]. Instead of non-locking implants used in a previous study
[41], the locking plate system

was adopted in the current study for posterior buttress plate fixation of the PLCF. The loading modes
in the test also required

further improvement. In a previous study [41], a synthetic distal femur was used to deliver axial
forces on total tibial plateau.

The uninjured, normal bone shared the majority of the load. Even though vertical displacement of
the posterolateral fragment

and integral failure load of each specimen were detected, the actual loading endured by the
posterolateral fragment could not

be identified or accurately measured. Therefore, it was decided that the loading should only be
applied on the posterolateral

fragment by a specially designed load-applicator, to obtain an explicit relationship between

concentrated loading and displace-

ment of the posterolateral fragment. The results show that the loading capacity endured by the
experimental models was signif-

icantly different from that reported in the literature [41,46].

Based on clinical facts, the level of loading also required adjustment.Whether the posterolateral
fragment needs such high fixation

strengths as posterior plate fixation was questioned. In accordance with previous rehabilitation
experience, patients who suffer from

tibial plateau fractures generally stay in bed and exercise with no weight loading of the knee in the
early postoperative stages. Knee

flexion (such as deep flexion N90°) with full weight bearing should not occur prior to union of the
fracture [47].Duringdeep flexion,

the maximal load on the knee joint is about four-to five-fold the patient's bodyweight, and there is
up to 70% larger peak pressure

in the medial compartment compared to the lateral [48]. Based on these considerations, the
ultimate load was set at 1000 N in the

current study. Clinically, the load on the posterolateral column of the tibial plateau is limited and
nomore than 1000 N. It is unlikely

that an extreme load on the posterolateral fragment would occur during the rehabilitation process.

In the current trauma center, the ‘magic screw’ technique has been applied to Schatzker type II
fractures with PLCF. There are

a few technical considerations in designing and using the ‘magic screw’. Firstly, the anteromedial
cortex of the proximal tibia is

tough, which could supply great anti-pull-out strength. Generally, a 3.5-mm locking screw is
selected, as shown in the current

study. The threaded tail of the locking screw is locked with the cortex. In addition, the medial aspect
of the tibia offers a wide

area for screw implantation. Secondly, instead of the traditional anterolateral to posterolateral
approach for screw implantation,

a longer screw could be implanted using the anteromedial to posterolateral approach, which could
provide stronger support to

the construct. Lastly, it does not interfere with lateral plate implantation, especially in a lateral
plateau fracture.

There is one more point that should not be ignored, themodified process of sample preparation for
testing (fixation in intact
synbone–implant removal–fracture modeling–fixation in situ–biomechanical test) in this
biomechanical study was different

from the traditional process (fracture modeling–fixation implantation–biomechanical test). During

the pilot study, the PLCF

fragment was easily broken by screwing or lost the reduction following the traditional process
because the fragment was

relatively small. Inspired by the process of olecranon osteotomy for distal humeral fracture and that
of malleolus osteotomy

for talus fracture, and by literature review [49,50], it was decided to try the new process in the
current study. The aim of

this new modified process was to facilitate anatomic reduction and optimum positioning of the
implant. In addition, there

was no impact on the biomechanical result after screw re-insertion in the current study, based on
findings from a previous

study [51]. So far, there is no published literature that compares these two processes, which was
also not the objective of

the current study. Throughout the study, the modeling process and approach were unified, which
assured that the different

group results were comparable.

There were some limitations in comparing this study to a clinical situation where the fracture feature
and the specimen quality

of donors are variable. For the different morphological types of PLCF, any of fixation patterns,
including the ‘magic screw’

technique, is appropriate for solving majority problems but not necessary for all of them; cases
should be carefully chosen for

this technique [29]. Compared with a depression fracture, modeling of a split-type fracture is
relatively simple. In the horizontal

plane, the depressed area of the posterolateral column, always located posterior to the anterior
border of the fibular head [52],

might have a great effect on the selection of the surgical approach and fixation pattern. For the
depression-type PLCF, further

design of fracture modeling and fixation simulation is required.

Human cadaveric bone, which is an ideal test material, was not used. However, a synthetic model
provides several advantages

over a cadaveric one. The synthetic tibiae provided standard dimensions and properties between
specimens, and thus, the implan-

tation techniques were reproducible and involved same size implant. Furthermore, the varying
dimensions, ages, and bone

mineral density were also avoided. Another weak point of this study was no cyclic loading test.
Therefore, the result of this

study must be interpreted as strictly static biomechanics, representing only part of the scenario at
work in the fixation, and no

healing of these injuries in vivo.

The results of the current study showed that although neither of the two lateral plate fixation
methods would have reached

the same strength as the posterior buttress plate fixation for PLCF, the combination fixation pattern
of a posteriorly positioned

lateral rafting plate with a ‘magic screw’ could provide significantly higher biomechanical stability
than a lateral plate alone,

and is similar to posterior plate fixation. With these findings, there may be reduced necessity and
usage of posterior plate fixation

through a posterior approach.