Injury, Int. J.

Care Injured 47 (2016) 617624

Contents lists available at ScienceDirect

Injury
journal homepage: www.elsevier.com/locate/injury

Inuence of plate material and screw design on stiffness and ultimate

load of locked plating in osteoporotic proximal humeral fractures
Jan Christoph Katthagen a,*, Michael Schwarze b, Mara Warnhoff a, Christine Voigt a,
Christof Hurschler b, Helmut Lill a
a
b

Department of Trauma and Reconstructive Surgery, Diakoniekrankenhaus Friederikenstift gGmbH, Humboldtstr. 5, 30169 Hannover, Germany
Laboratory of Biomechanics and Biomaterials, Medizinische Hochschule Hannover (MHH), Anna-von-Borries-Str. 1-7, 30625 Hannover, Germany

A R T I C L E I N F O

A B S T R A C T

Article history:
Accepted 9 January 2016

Introduction: The main purpose was to compare the biomechanical properties of a carbon-bre
reinforced polyetheretherketone (CF-PEEK) composite locking platewith pre-existing data of a titaniumalloy plate when used for xation of an unstable 2-part fracture of the surgical neck of the humerus. The
secondary purpose was to compare the mechanical behaviour of locking bolts and conventional locking
cancellous screws.
Methods: 7 pairs of fresh frozen human humeri were allocated to two equal groups. All specimens were
xed with the CF-PEEK plate. Cancellous screws (PEEK/screw) were compared to locking bolts (PEEK/
bolt) for humeral head xation. Stiffness, fracture gap deection and ultimate load as well as load before
screw perforation of the articular surface were assessed. Results were compared between groups and
with pre-existing biomechanical data of a titanium-alloy plate.
Results: The CF-PEEK plate featured signicantly lower stiffness compared to the titanium-alloy plate
(P < 0.001). In ultimate load testing, 6 out of 14 CF-PEEK plates failed due to irreversible deformation and
cracking. No signicant difference was observed between results of groups PEEK/screw and PEEK/bolt
(P > 0.05).
Discussion: The CF-PEEK plate has more elastic properties and signicantly increases movement at the
fracture site of an unstable proximal humeral fracture model compared to the commonly used titaniumalloy plate. The screw design however does neither affect the constructs primary mechanical behaviour
in the constellation tested nor the load before screw perforation.
2016 Elsevier Ltd. All rights reserved.

Keywords:
Carbon bre
PEEK
Proximal humerus fracture
Locked plating
Screw

Introduction
Locking-screw plating is an established and commonly used
treatment for dislocated proximal humeral fractures [14]. Despite
advancement of implants and a better understanding of failure
mechanisms, complication rates for proximal humeral locked plate
xation remain high at up to 30% of cases being reported [13,5]. A
large number of complications and unsatisfactory treatment
results can be attributed to varus loss of reduction and articular
screw perforation [13,68]. Although locked plating has improved the functional outcome, complications are partially
ascribed to the rigidity of locked plating in osteoporotic bone
[2,3]. Newly developed implants are regularly introduced to the

* Corresponding author. Tel.: +49 511 2330; fax: +49 511 2405.
E-mail address: christoph.katthagen@ddh-gruppe.de (J.C. Katthagen).
http://dx.doi.org/10.1016/j.injury.2016.01.004
00201383/ 2016 Elsevier Ltd. All rights reserved.

market which attempt to address these issues with the objective of

reducing complications and improving treatment results.
Locking plates made of carbon bre (CF) reinforced polyetheretherketone (PEEK) composite are one example of these
innovations. CF-PEEK composite plates are less stiff than
stainless steel and titanium plates, with an elastic modulus
similar to bone as well as the ability to withstand prolonged
fatigue strain [911]. Reduced implant stiffness in proximal
humeral fracture treatment was proven to unload the bonescrew tip interface which might lead to a reduction of implant
associated complications [1214]. The development of CF-PEEK
plates for xation of proximal humeral fractures derived from the
idea to combine the advantages of locked plating with the
potential advantages of reduced implant rigidity. A further
benet of CF-PEEK is that the radiolucency of polymer
composites allows for good imaging of fracture healing and
early evaluation of possible complications [15].

618

J.C. Katthagen et al. / Injury, Int. J. Care Injured 47 (2016) 617624

New designs of locking screws are another innovative development. Stable screw anchorage within the osteoporotic cancellous
bone of the proximal humerus is a major challenge of successful
fracture treatment in the typical patient population [16]. Screws
with a thicker diameter and a lower thread depth compared to
conventional screws are intended to provide a larger loading surface
within the cancellous bone [17]. Furthermore, a rounded rather than
a pointed screw tip is intended to better underpin the far cortex with
lower rate of articular screw perforation when used for proximal
humeral fracture xation [17]. At this point, it is unknown whether
the differences in screw design affect overall mechanical behaviour
of a proximal humeral fracture model.
The purpose of this study was to compare the biomechanical
properties of a CF-PEEK composite proximal humeral locking plate
with pre-existing data of a titanium-alloy plate when used for
xation of an unstable 2-part fracture of the surgical neck of the
humerus. The secondary purpose was to investigate the mechanical
performance and screw perforation behaviour of newly developed
locking bolts in comparison to conventional locking humeral head
cancellous screws. We hypothesised that the CF-PEEK plate would
feature less stiffness compared to the titanium-alloy plate. Concerning the new locking bolts, we expected reduced fracture gap motion
when compared to conventional locking humeral head cancellous
screws. Additionally, we hypothesised an increased load bearing for
the locking bolts before cut out through the articular surface.

Fig. 1. Material used in the tests: A: Left 4.0 mm cancellous locking screw; Right
4.0 mm locking bolt with low thread depth and rounded screw tip (Arthrex1,
Karlsfeld, Germany); B: Left - CF-PEEK proximal humeral plate (PEEK Power
Humeral Fracture Plate; Arthrex1, Karlsfeld, Germany); Right: titanium-alloy plate
(PHILOS; DePuy Synthes1, Umkirch, Germany).

Material and methods

Specimens
Seven matched pairs of fresh-frozen humeri (n = 14) attained
from female human donors (mean age 59 years; range, 4675
years) were randomised into two groups. Bony pathologies were
ruled out by means of radiographs taken in two planes
(anteroposterior and axial). The bone mineral density (BMD) of
each humeral head was assessed with dual X-ray absorptiometry
(Discovery QDR Series; Hologic1, Bedford, MA, USA).
Implants and groups
The carbon-bre reinforced PEEK proximal humeral locking
plate (PEEK Power Humeral Fracture Plate; Arthrex1, Karlsfeld,
Germany) was used in 90 mm length (Fig. 1). This plate is

manufactured from continuous carbon bres (5560%) and

injection-moulded PEEK and is 2.6 mm thick. The assembly of
the carbon bres is multidirectional and randomly oriented.
In the PEEK/screw group, seven conventional 4.0 mm cancellous locking screws (Figs. 1 and 2; Cancellous Screw, Arthrex1,
Karlsfeld, Germany) were used to x the humeral head. In the
second group (PEEK/bolt), seven newly developed 4.0 mm locking
bolts (Locking bolt, Arthrex1, Karlsfeld, Germany) with low thread
depth and rounded screw tip (Figs. 1 and 2) were used to stabilise
the proximal humerus. In both groups three cortical screws were
used to stabilise the shaft. All screws were made of titanium alloy
(ASTM F136; Ti6Al4V; Arthrex1, Karlsfeld, Germany). The lengths
of all screws were not predened but selected such that their tips
extended to the subchondral surface of the humeral head without
penetration of the articular surface. Locking screw xation was
achieved by threading the screw heads into the CF-PEEK-plate. A

Fig. 2. X-rays in two planes of a left humerus after osteotomy and xation with the CF-PEEK plate and seven conventional. A: 4.0 mm cancellous locking screws in the humeral
head. B: Newly developed 4.0 mm locking bolts in the humeral head.

J.C. Katthagen et al. / Injury, Int. J. Care Injured 47 (2016) 617624

jig guided the screws directions and screw insertion was stopped
once the screws head was ush with the plate. Insertion torque
was not controlled.
The results of this investigation were compared with data from
a previous biomechanical study performed in our lab that
evaluated the PHILOS proximal humerus xation plate (DePuy
Synthes1, Umkirch, Germany), with exactly the same biomechanical test setup, fracture model and testing mode [18]. The PHILOSplate is made of titanium alloy; the short plate has a length of
90 mm, a thickness of 2.8 mm and a comparable shape as the CFPEEK plate (Fig. 1). The data in the previous study was attained
using eight female human humeri with similar bone properties
(BMD 0.43(0.12) g/cm2). With eight 3.5 mm locking screws in the
humeral head and three screws along the shaft, screw conguration was similar as tested in the study presented herein (Fig. 3).
Locking bolts of PEEK/bolt resembled the locking screws used with
the titanium-alloy plate more than the cancellous screws of PEEK/
screw. Therefore results of PEEK/bolt were compared to the results
of the previous study.
Specimen preparation
As with the methodology of the previously published investigation including the PHILOS-plate [18], the distal part of the
humerus was cut 22 cm from the most proximal point of the
humeral head. The long axis of the bone was aligned vertically in
the coronal and sagittal planes and the distal 4 cm of the remaining
humeral shaft was potted into an aluminium cylinder with a coldcuring casting resin (Rencast FC 53, Huntsman Advanced Materials,
Basel, Switzerland). In order to simulate an unstable 2-part
fracture of the proximal humerus with comminuted medial cortex,
a standardised transverse wedge osteotomy was created using a
custom made jig. The gap osteotomy was positioned in relation to
the CF-PEEK plate. The rst cut was placed directly below the
plate-hole for the calcar screw (most distal head hole); the second
cut was 10 mm above, leaving a 10 mm gap at the region of the
surgical neck.
Stiffness tests
The entire test setup, testing modes and sequence of tests was
identical with the methodology of a previously published
investigation [18]. Briey, construct stiffness was assessed during

619

torsional and axial loading as well as axial loading with the

construct abducted and adducted by 208. Fracture gap deection
was assessed during cyclic loading. Finally all specimens were
loaded to ultimate failure in axial direction, simulating a fall
scenario on the outstretched arm. All stiffness and strength tests
were done on a MiniBionix 858 (MTS, Eden Prairie, MN, USA); a
linear actuator with a custom control program was used to load the
humeral heads.
Torsional tests were chosen to imitate internal and external
rotation of the shoulder. For torsion stiffness tests the distal potted
part of each specimen was mounted on a universal joint (Fig. 4).
The humeral head was positioned centrally inside a cylindrical
aluminium chamber and xed with ten blunt screws on two levels
leaving the plate spared. The proximal cylinder was mounted to
another universal joint connected with the actuator. Torsional load
was applied to the humeral head at a constant speed of 0.18/s with
a load limit of + and 3.5 Nm. Displacement of the humeral head
was measured grip to grip with the MiniBionix. To account for
the partly non-monotonically increasing moment-displacement
correlation, secant stiffness for each specimen was calculated.
Torsional stiffness was assessed three times each for external and
internal rotation of the actuator after one preconditioning cycle;
the mean values were calculated and analysed.
For the following stiffness tests, cyclic loading and failure test,
fracture gap displacement was determined three dimensionally by
an ultra-sound based device with an accuracy of 0.1 mm (CMS 20S,
Zebris Medical, Germany). The distal potted part of the humerus
was mounted to a tilting block which was left in horizontal
position for axial stiffness test (Fig. 4). A custom made dynamic
plate was mounted to the linear actuator to eliminate shear forces
and only transmit axial forces during loading. One preconditioning
cycle with a load of 200 N was performed. Vertical, non-destructive
load was then applied to the apex of the humeral head at a constant
speed of 0.1 mm/s to a load limit of 200 N in, 208 abduction, and 208
adduction respectively. The axial testing mode was chosen
following the basic test setup of Lescheid et al. [19]. Abduction
mode was selected to simulate shear loading across the proximal
fracture site as experienced during rising out of a chair or crutch
weight bearing [19]. The adduction setup was orientated on the
results of glenohumeral load transmission of in vivo measurements [20,21]. To account for the partly non-monotonically
increasing force-displacement correlation, the secant stiffness
for each specimen was calculated. All tests were repeated three

Fig. 3. Lineup of different xations compared: A: Specimen of group PEEK/bolt xed with the CF-PEEK plate and seven 4.0 mm locking bolts in the humeral head. B: Specimen
of a previous study [18] xed with the PHILOS plate and 8 conventional locking screws around the humeral head.

620

J.C. Katthagen et al. / Injury, Int. J. Care Injured 47 (2016) 617624

Fig. 4. A: Setup for stiffness tests in axial mode, cyclic testing and load to failure mode: the distal potted part is rigidly xed to the tilting block (asterisk), the dynamic plate is
connected to the actuator (#). Displacement at the fracture gap is determined three dimensionally by an ultra-sound based device (arrows). B: Setup for stiffness tests in
torsion mode: the distal and proximal parts are xed on universal joints (). C: Setup for screw cut out testing: A PMMA cylinder with a circular cavity is loading the rigidly
xed humeral head.

times for each specimen, and the average stiffness of the three
loading cycles calculated.

the screw tips. The test was stopped once screw perforation with
exposed screw tip and/or cartilage tear was observed visually.

Cyclic and failure loading

Statistics

Stiffness tests were followed by cyclic axial loading.

5000 cycles of 50250 N load were applied at 1 Hz in a sinusoidal
waveform, which corresponds to the number of arm movements
within one week of postoperative mobilisation [22]. Displacement at the fracture gap was continuously monitored. Mean
displacement for the entire 5000 cycle testing period and
maximum displacement at cycle number 5, 50, 100, 2500 and
5000 were calculated. Finally all specimens were loaded to failure
applying vertical force at 0.1 mm/s mimicking the loading that
might occur during a fall from standing or during postoperative
full loading. The plates were visually checked after each series of
loads so that the time point of failure could be determined. Failure
was dened as one of the following: (1) gap closure, dened as
contact between the medial cortices; (2) fracture around the
humeral head or shaft; (3) implant failure (e.g. cracking of the
plate).

Normal distribution could not be assumed and was tested with

the KolmogorovSmirnovTest. The WilcoxonMannWhitneytest was performed to compare stiffness of each loading mode,
load-to-failure and screw perforation tests between groups PEEK/
screw and PEEK/bolt (n = 7 each). The multivariate PillaiSpur-Test
was used to tests for signicant difference between PEEK/screw
and PEEK/bolt during cyclic loading. In order to compare
biomechanics of the CF-PEEK plate and of the PHILOS plate,
results of PEEK/bolt were compared with results from a study using
the same testing protocol [18] using the Wilcoxon-Mann
Whitney-test. The signicance level alpha was set to 0.05 for all
comparisons.

Screw perforation
After the previous tests, the humeral shaft was cut beneath the
plate and an additional 8 mm gap osteotomy was placed at the
anatomic neck of the humeral head. The plate-screw xation of
each humeral head was left unchanged and the construct was
potted in a brass ring, with the articular surface facing upwards.
The bone-implant-interface was xed in casting resin so that prior
plate cracking did not affect the screws stability within the plate.
Similar to an already published setup, a transparent PMMA
cylinder with a circular cavity tting the articular surface was
mounted under the dynamic plate, which eliminated shear forces
(Fig. 4C) [23]. Load was applied with constant displacement of
0.05 mm/s pressing the humeral head articular fragment against

Results
Mean neck-shaft-angle of all specimens was 134.7(2.4)8. With
an average of 0.42(0.13) g/cm2 BMD, the humeral heads were
osteoporotic [24,25]. BMD of the specimens did not differ between
the groups compared.
CF-PEEK versus PHILOS
Results of PEEK/bolt were compared with results from testing of
the PHILOS titanium plate from a previous study [18] (Fig. 3). The
CF-PEEK plate demonstrated signicantly lower stiffness in all
tests (Table 1; internal rotation, P < 0.001; external rotation,
P < 0.001; axial stiffness, P < 0.001; abduction, P < 0.001; adduction P < 0.001) compared to results of the PHILOS plate (Fig. 5).
During cyclic loading, signicant differences between fracture gap
deection of both groups was observed (P = 0.001; Fig. 6).

J.C. Katthagen et al. / Injury, Int. J. Care Injured 47 (2016) 617624

621

Table 1
Results of the CF-PEEK plate with locking bolts and the titanium alloy plate with
locking screws: mean stiffness with standard deviation of all testing modes and
mean load to failure with standard deviation (brackets).
CF-PEEK with
locking bolts
(n = 7)
Internal rotation
(Nm/8)
External rotation
(Nm/8)
Axial stiffness
(N/mm)
Abduction
(N/mm)
Adduction
(N/mm)
Failure load
(N)
*

Titanium-alloy
with locking
screws (n = 8)

Signicance

Power

0.29 (0.06)

0.42 (0.48)

P < 0.001*

0.10

0.28 (0.06)

0.59 (0.15)

P < 0.001*

0.99

P < 0.001

1.00

0.99

57.1 (15.5)

183.7 (26.3)

46.1 (11.8)

118.4 (32.2)

P < 0.001

96.1 (29.9)

505.5 (321.2)

P < 0.001*

0.88

822 (143)

P = 0.69

0.05

801 (201)

Signicance between results of the CF-PEEK plate and the titanium-alloy plate.

No signicant difference (P = 0.69) was observed regarding

ultimate load. In group PEEK/bolt failure mode was gap closure by
plate bending in 3cases (3/7), implant cracking in 3 cases (3/7),
and shaft facture in one case (1/7). The PHILOS plate had failed
by gap closure due to plate bending in six cases (6/8) and by
humeral shaft fracture in two cases (2/8).
Screws versus bolts
No signicant difference could be observed between stiffness of
PEEK/screw compared to PEEK/bolt (Table 2; internal rotation,
P = 0.26; external rotation, P = 0.8; abduction, P = 0.8; adduction,
P = 0.26; axial stiffness, P = 0.54). Subsequent post hoc power
analysis indicated small effect sizes (Table 2). Fracture gap
deection of both groups during cyclic loading showed similar
curve progression and no signicant difference was observed
(Fig. 6; P = 0.72). Furthermore, no signicant difference could be
detected between failure loads of both groups (Table 2; P = 0.32).
In group PEEK/screw, failure mode was gap closure by plate
bending in 4 cases (4/7) and implant cracking in 3 cases (3/7). In
group PEEK/bolt failure mode was gap closure by plate bending in

Fig. 6. Deection of CF-PEEK plate under cycling loading with cancellous screws
(red) and locking bolts (green). For comparison the results from the titanium-alloy
plate with similar screw conguration are integrated (blue). Error bars indicate the
95% condence intervals for each curve.

3 cases (3/7), implant cracking in 3 cases (3/7), and shaft facture in

one case (1/7). In all cases construct failure occurred during
monotonic loading to failure. The PEEK implant broke in all cases of
both groups (6/14) by small horizontal cracks forming along the
distal humeral head screws and the most distal suture eyelets
(Fig. 7).
Screw perforation
Screw perforation was observed at a mean load of
851(SD:385) N for the PEEK/screw group and 933(SD:449) N
for the PEEK/bolt group, differences were not signicant
(P = 0.96, Fig. 8). Perforating screws were the calcar (PEEK/
screw n = 3, PEEK/bolt n = 2), the anterior-medial (PEEK/screw
n = 0, PEEK/bolt n = 4), the anterior-superior (PEEK/screw n = 2,
PEEK/bolt n = 1), the posterior-inferior (PEEK/screw n = 1, PEEK/
bolt n = 0) and the anterior-inferior (PEEK/screw n = 1, PEEK/bolt
n = 0) screw. In group PEEK/screw the sharp screw tips

Table 2
Mean stiffness and standard deviation of all groups and testing modes. Mean load to
failure and standard deviation (in brackets) of all groups.
Group
PEEK/screw
n=7

Fig. 5. Results of stiffness tests of the CF-PEEK-plate with locking bolts (PEEK/bolt)
compared with results of the PHILOS plate with locking screws in axial, adduction
and abduction modes. Brackets with asterisks indicate signicant differences
between results. Single dots represent outliers with deviation of more than
1.5 * inter-quartile range from the median.

Internal rotation
(Nm/8)
External rotation
(Nm/8)
Axial stiffness
(N/mm)
Abduction
(N/mm)
Adduction
(N/mm)
Failure load
(N)

Group
PEEK/bolt
n=7

Signicance

Power

0.26 (0.06)

0.29 (0.06)

P = 0.26

0.13

0.27 (0.06)

0.28 (0.06)

P = 0.8

0.06

54.7 (18.8)

57.1 (15.5)

P = 0.54

0.06

44.6 (7.5)

46.1 (11.8)

P = 0.8

0.06

82.2 (18.1)

96.1 (29.9)

P = 0.26

0.16

653 (133.5)

800.9 (201)

P = 0.32

0.30

622

J.C. Katthagen et al. / Injury, Int. J. Care Injured 47 (2016) 617624

Fig. 7. Example of implant failure during ultimate loading. Small cracks can be
observed along the distal humeral head screw holes and the distal suture eyelets
(arrows).

perforated the surface with exposed screw tip whereas in group

Peek/bolt a bulge of the articular surface with tear of the
cartilage layer was observed (Fig. 9).
Discussion
The purpose of this study was to compare the biomechanical
properties of a CF-PEEK composite proximal humeral locking plate

Fig. 8. Load applied when rst screw perforation was observed dependent on
screws or bolts.

with pre-existing data of a titanium-alloy plate when used for

xation of an unstable 2-part fracture of the surgical neck of the
humerus. Secondary purpose was to investigate the mechanical
behaviour of newly developed locking screws in comparison to
conventional locking humeral head cancellous screws.
As hypothesised, the CF-PEEK plate demonstrated signicantly
lower stiffness compared to a commonly used titanium-alloy plate.
6 out of 14 CF-PEEK plates failed due to irreversible deformation
and cracking during ultimate loading. The cracks of the CF-PEEK
resulted in a permanent loss of xation stability. The new locking
bolts did not reduce fracture gap deection of the unstable 2-part
proximal humeral fracture model and did not increase load before
screw perforation compared to cancellous locking screws.
This time zero study evaluates the mechanical behaviour of a
new xation device and provides baseline information of the
mechanical integrity between constructs at the time of surgery. As
a major limitation to our ndings, the results only represent initial
xation and do not reect biological processes of in vivo healing or
the loading produced by muscle forces acting around the shoulder
[26]. Especially the traction of the rotator cuff deemed responsible
for varus loss of reduction was not simulated. However, as early
rehabilitation after locked plating of proximal humeral fractures
often begins with active-assisted joint mobilisation, the mechanical in vivo conditions within the rst postoperative week are
simulated [2,4,5]. Furthermore, direction of load application and
forces transmitted through the fracture site were within the range
of glenohumeral contact forces measured in vivo during activities
of daily living [20,21,27]. Specimens used in this investigation
represent the typical patient population of elderly females with
osteoporotic bone structure [13,6,18,24,25]. The 2-part fracture
model used in this study is common in biomechanical testing of
proximal humerus fractures, allows good reproducibility and
represents the second most common fracture type treated by
locked plating [1,3,11,19,22,26,28]. However, more complex 3- and
4-part fractures are not represented. Results of the CF-PEEK plate
were compared to pre-existing data of a titanium-alloy plate
resulting from a previous study using exactly the same setup and
testing modes. Cadaver humerus assignment was not directly
matched between the studies, but BMD of the humeral heads did
not differ between the groups compared. Further potential bias,
e.g. by anthropometric differences, was neglected.
Compared to the current gold standard of locked plate xation,
the CF-PEEK plate showed more elastic properties and signicantly
increased movement at the fracture site of an unstable proximal
humeral fracture model. A previous biomechanical study determined in vitro characteristics of ve different implants for the
stabilisation of proximal humeral fractures under static and cyclic
loading [13]. The authors found that a sufciently stable
osteosynthesisexible enough to unload the implant-bone
interface and rigid enough to minimise fracture movementswill
prove successful in clinical use. Within this context our ndings
raise two remaining questions: (1) Might bone healing be impaired
by the low plate rigidity and the consequently resulting large
movements at the fracture site? (2) Could the reduced stiffness of
the plate lead to unloading of the screw-tip bone interface with
consequent decrease of articular screw perforations especially in
osteoporotic bone?
The torsional stiffness of the construct was low enough that the,
3.5 Nm torsional loads resulted in 12138 of relative rotation
between the proximal and distal fragments. Similarly, the axial
stiffness was low enough that the 200 N applied loads resulted in
interfragmentary displacements of 4 mm. These strains are
sufciently high that they could conceivably have a signicant
negative impact on bone healing especially in case of large defect
zones [29,30]. The upper limit of stability required for successful
fracture healing depends upon many factors und basically remains

J.C. Katthagen et al. / Injury, Int. J. Care Injured 47 (2016) 617624

623

Fig. 9. Typical perforation types of screws after the loading: A: sharp screw tip perforating the surface in group Peek/screw B: Bulge with partial tear of cartilage layer in the
Peek/bolt group.

unknown for each specic fracture situation [30]. Although larger

interfragmentary movements and strains lead to larger callus
formation this effect is limited with increasing fracture gap size
[29]. Against this background the clinical use of CF-PEEK plates should
be cautious in case of remaining large fracture gaps of more than 4
5 mm after fracture reduction without additional means of fracture
gap augmentation.
The discussion of question (2) is more speculative. Glenohumeral contact forces act on the region of the humeral head in which
the mono-cortical xed humeral head screws end [20,21,27]
(Figs. 2 and 3). The humeral heads joint surface is loaded against the
screw-tips during shoulder motion. In a previous biomechanical
investigation concerning how stiff an implant should be for
proximal humeral fracture xation the following was concluded:
implants with low stiffness and elastic characteristics. . .appear to
minimise peak stresses at the bone implant interface, making them
particularly suitable for fracture xation in osteoporotic bone
[11]. Within this context is might be assumed that the reduced
stiffness of CF-PEEK plates could unload the bone-screw tip
interface to a certain extent which can be considered benecial
in particular in osteoporotic bone [13,14]. This hypothesis is
supported by recent ndings of a computed nite element
simulation of a three part proximal humeral fracture [12]. The
authors found that devices manufactured from short bre CF-PEEK
will signicantly reduce cortical bone stresses in the region of the
screw tips, and may provide lower instances of device failure by
screw pullout/pushout. In particular in the setting of osteoporotic
proximal humeral fractures of the elderly, locked CF-PEEK plates
might therefore contribute to the reduction of articular screw cutout by means of their reduced stiffness, which itself reduces the
load on the bone- screw tip interface. Both raised questions
following the ndings of this study demand for further investigation. At this point it should be mentioned that the relevance of
biocompatibility, which has historically been an issue with carbon
bre devices that produce wear debris, has not yet been fully
claried [9]. With respect to recent ndings however, biocompatibility might no more be a major concern with novel implants [8,13].
The high rate of implant failure resulting from cracking of the
CF-PEEK plates during ultimate loading in this investigation (6/14)
must be considered a serious drawback. Generally CF composite
materials have little tolerance for plastic deformation and failure
tests are regularly terminated by a resultant plate breakage
[9,10]. The investigated facture model depicts a worst case scenario
with large defect zone which does not exist in all fracture types. The
mean failure load of the CF-PEEK plate did not signicantly differ
from the mean failure load of the PHILOS plate, which itself showed

no implant breakage during biomechanical testing but during

clinical use [2]. Generally failure load of CF-PEEK composite plates
can be increased with increased implant thickness [9]. Modication
and thickening of the PEEK Power Humeral Fracture Plate
(Arthrex1, Karlsfeld, Germany) in the region of implant failure
might possibly lower the risk of implant failure in clinical practice.
The secondary hypothesis that locking bolts would reduce
fracture gap deection compared to cancellous locking screws was
not conrmed by ndings in this study. The geometry of locking
bolts with their low thread depth is more voluminous and is
assumed to better support the humeral head against varus
displacement compared to cancellous locking screws (Fig. 1). Still,
none of the tests showed any signicant difference between the
groups. Generally, the assumed effect by locking bolts might be
only marginal and cannot be observed in conjunction with the
predominant low stiffness of the CF-PEEK plate in context of an
unstable fracture model.
The desired clinical effect of the newly developed locking bolts
with their large loading surface and the rounded screw tip is to
reduce the rate of articular screw cut-out in osteoporotic bone
[17]. Specimens in this investigation were cyclically loaded
5000 times, which corresponds to the number of arm movements
within one week of postoperative mobilisation [22]. Nonetheless,
no articular screw perforation was observed in any of the specimens
of either group tested during cyclic loading. Unfortunately, up to
date there is no biomechanical bench test that was proven to
reliably reproduce the clinically typical failure mode with varus
displacement and articular screw perforation. In a previous
biomechanical investigation about the inuence of numbers and
position of screws on perforation of the humeral head, the load on
the articular surface was stepwise increased until one or more
screws perforated the articular surface [23]. Within our comparable
test setup for investigation of screw perforation, the humeral heads
joint surface was continuously pressed against the underlying
screws tips. Other than expected, the larger loading surface and the
rounded screw tips of the locking bolts did not increase resilience
against articular perforation. However, we observed that when
conventional screws perforated the surface, a small and sharp tip
was visible. During perforation of the locking bolts, the tip was still
covered by a layer of cartilage (Fig. 9), which could prevent
secondary destruction of the glenoid in clinical practice.
Conclusions
The CF-PEEK plate has more elastic properties and signicantly
increases movement at the fracture site of an unstable proximal

624

J.C. Katthagen et al. / Injury, Int. J. Care Injured 47 (2016) 617624

humeral fracture model compared to the commonly used

titanium-alloy plate. The screw design, however, does neither
affect the constructs primary mechanical behaviour in the
constellation tested nor the load before screw perforation.
Conict of interest statement
J. Christoph Katthagen, Michael Schwarze, Mara Warnhoff,
Christine Voigt, and Christof Hurschler This author, their
immediate family, and any research foundation with which they
are afliated did not receive any nancial payments or other
benets from any commercial entity related to the subject of this
article.
Helmut Lill is consultant with Arthrex1, Karlsfeld, Germany
and DePuy Synthes1, Umkirch, Germany. However, the author and
his immediate family, and any research foundation with which
they are afliated did not receive any nancial payments or other
benets from any commercial entity related to the subject of this
article.
Acknowledgements
This study was funded by the Alwin Jager foundation. All
implants used in this investigation were provided by Arthrex1
(Karlsfeld, Germany). The study sponsors had no involvement in
the study design, in the collection, analysis and interpretation of
data, in the writing of the manuscript and in the decision to submit
the manuscript for publication. The senior author is consultant
with Arthrex1, Karlsfeld, Germany and DePuy Synthes1, Umkirch,
Germany.
References
[1] Sproul RC, Iyengar JJ, Devcic Z, Feeley BT. A systematic review of locking plate
xation of proximal humerus fractures. Injury 2011;42:40813.
[2] Sudkamp N, Bayer J, Hepp P, Voigt C, Oestern H, Kaab M, et al. Open reduction and
internal xation of proximal humeral fractures with use of the locking proximal
humerus plate. Results of a prospective, multicenter, observational study. J Bone
Jt Surg Am 2009;91:13208.
[3] Thanasas C, Kontakis G, Angoules A, Limb D, Giannoudis P. Treatment of
proximal humerus fractures with locking plates: a systematic review. J Shoulder Elbow Surg 2009;18:83744.
[4] Siebenburger G, Biermann N, Haasters F, Mutschler W, Ockert B. Klinisches
Outcome (Median 5 Jahre) nach winkelstabiler Plattenosteosynthese von
228 Patienten mit dislozierter proximaler Humerusfraktur. Obere Extrem
2014;9:21521.
[5] Hillbricht S, Sturzebecher K, Suhren J, Kasch R, Herzog J, Schulz AP. Plate
osteosynthesis or intramedullary nailing for proximal humerus fracture
retrospective analysis of 156 cases over 3 years with a minimum 2 years
follow. Obere Extrem 2013;8:144.
[6] Hardeman F, Bollars P, Donnelly M, Bellemans J, Nijs S. Predictive factors for
functional outcome and failure in angular stable osteosynthesis of the proximal humerus. Injury 2012;43:1538.
[7] Schliemann B, Wahnert D, Theisen C, Herbort M, Kosters C, Raschke MJ, et al.
How to enhance the stability of locking plate xation of proximal humerus
fractures? An overview of current biomechanical and clinical data. Injury
2015;46:120714.
[8] Erasmo R, Guerra G, Guerra L. Fractures and fracture-dislocations of the
proximal humerus: a retrospective analysis of 82 cases treated with the
Philos1 locking plate. Injury 2014;45:S438.

[9] Fujihara K, Huang Z-M, Ramakrishna S, Satknanantham K, Hamada H. Performance study of braided carbon/PEEK composite compression bone plates.
Biomaterials 2003;24:26617.
[10] Steinberg EL, Rath E, Shlaifer A, Chechik O, Maman E, Salai M. Carbon ber
reinforced PEEK Optima a composite material biomechanical properties and
wear/debris characteristics of CF-PEEK composites for orthopedic trauma
implants. J Mech Behav Biomed Mater 2013;17:2218.
[11] Kurtz SM, Devine JN. PEEK biomaterials in trauma, orthopedic, and spinal
implants. Biomaterials 2007;28:484569.
[12] Feerick EM, Kennedy J, Mullett H, FitzPatrick D, McGarry P. Investigation of
metallic and carbon bre PEEK fracture xation devices for three-part proximal humeral fractures. Med Eng Phys 2013;35:71222.
[13] Lill H, Hepp P, Korner J, Kassi JP, Verheyden AP, Josten C, et al. Proximal
humeral fractures: how stiff should an implant be? A comparative mechanical
study with new implants in human specimens. Arch Orthop Trauma Surg
2003;123:7481.
[14] Hessmann MH, Sternstein W, Mehler D, Korner J, Hofmann A, Rommens PM.
Are angle-xed implants with elastic properties advantageous for the internal
xation of proximal humerus fractures? Biomed Tech (Berl) 2004;49:34550.
[15] Cotic M, Vogt S, Hinterwimmer S, Feucht MJ, Slotta-Huspenina J, Schuster T,
et al. A matched-pair comparison of two different locking plates for valgusproducing medial open-wedge high tibial osteotomy: peek-carbon composite
plate versus titanium plate. Knee Surg Sports Traumatol Arthrosc 2014. http://
dx.doi.org/10.1007/s00167-014-2914-8.
[16] Lill H, Ellwein A, Katthagen C, Voigt C. Osteoporotic fractures of the proximal
humerus. Chirurg 2012;83:85865.
[17] Freude T, Schroeter S, Plecko M, Bahrs C, Martetschlaeger F, Kraus TM, et al.
Dynamic-locking-screw (DLS)-leads to less secondary screw perforations in
proximal humerus fractures. BMC Musculoskelet Disord 2014;15:194.
[18] Katthagen JC, Schwarze M, Meyer-Kobbe J, Voigt C, Hurschler C, Lill H.
Biomechanical effects of calcar screws and bone block augmentation on
medial support in locked plating of proximal humeral fractures. Clin Biomech
(Bristol Avon) 2014;29:73541.
[19] Lescheid J, Zdero R, Shah S, Kuzyk PRT, Schemitsch EH. The biomechanics of
locked plating for repairing proximal humerus fractures with or without
medial cortical support. J Trauma 2010;69:123542.
[20] Westerhoff P, Graichen F, Bender A, Halder A, Beier A, Rohlmann A, et al. In vivo
measurement of shoulder joint loads during activities of daily living. J Biomech
2009;42:18409.
[21] Bergmann G, Graichen F, Bender A, Rohlmann A, Halder A, Beier A, et al. In vivo
gleno-humeral joint loads during forward exion and abduction. J Biomech
2011;44:154352.
[22] Chow RM, Begum F, Beaupre LA, Carey JP, Adeeb S, Bouliane MJ. Proximal
humeral fracture xation: locking plate construct  intramedullary bular
allograft. J Shoulder Elb Surg 2012;21:894901.
[23] Erhardt JB, Stoffel K, Kampshoff J, Badur N, Yates P, Kuster MS. The position and
number of screws inuence screw perforation of the humeral head in modern
locking plates: a cadaver study. J Orthop Trauma 2012;26:e18892.
[24] Lill H, Hepp P, Gowin W, Oestmann JW, Korner J, Haas NP, et al. Age- and
gender-related distribution of bone mineral density and mechanical properties of the proximal humerus. Rofo 2002;174:154450.
[25] Tingart MJ, Apreleva M, von Stechow D, Zurakowski D, Warner JJ. The cortical
thickness of the proximal humeral diaphysis predicts bone mineral density of
the proximal humerus. J Bone Jt Surg Br 2003;85:6117.
[26] Bae J-H, Oh JK, Chon CS, Oh CW, Hwang JH, Yoon YC. The biomechanical
performance of locking plate xation with intramedullary bular strut graft
augmentation in the treatment of unstable fractures of the proximal humerus.
J Bone Jt Surg Br 2011;93:93741.
[27] Bergmann G, Graichen F, Bender A, Kaab M, Rohlmann A, Westerhoff P. In vivo
glenohumeral contact forces measurements in the rst patient 7 months
postoperatively. J Biomech 2007;40:213949.
[28] Mathison C, Chaudhary R, Beaupre L, Reynolds M, Adeeb S, Bouliane M.
Biomechanical analysis of proximal humeral xation using locking plate
xation with an intramedullary bular allograft. Clin Biomech (Bristol Avon)
2010;25:6426.
[29] Claes L, Augat P, Suger G, Wilke HJ. Inuence of size and stability of the
osteotomy gap on the success of fracture healing. J Orthop Res
1997;15:57784.
[30] Schell H, Epari DR, Kassi JP, Bragulla H, Bail HJ, Duda GN. The course of bone
healing is inuenced by the initial shear xation stability. J Orthop Res
2005;23:10228.