Keywords screw loosening - screw failure - fracture healing - ovine - screw retention
Introduction
Screws are the most commonly used implant in orthopaedic surgery.[1 ] However, screws do not always operate effectively and may loosen leading to fracture
non-union.[2 ]
[3 ] Screw failure is especially prevalent following procedures that treat patients with
poor bone quality,[3 ]
[4 ]
[5 ] and is one of the primary causes of revision procedures for orthopaedic hardware.[4 ] These revision procedures typically require extensive preoperative planning, the
use of specialized implants and tools and mastery of technically challenging surgical
techniques that dramatically raise health care costs.[6 ] There is a wide range of screw failure processes that can include screw backout,
stripping, complete fracture or loosening due to infection.[3 ]
[7 ] Screw loosening is considered an unsolved issue and new techniques or devices that
prevent loosening would represent a significant development for orthopaedic clinical
practice.
To reduce the incidence of screw loosening, various groups have focused on improving
screw/plate technologies,[8 ]
[9 ]
[10 ]
[11 ] among other approaches including bone cements, materials from the operating room
and high friction surface coatings.[12 ]
[13 ]
[14 ] Many of these efforts are mechanical solutions implemented intraoperatively and
carry additional sets of risks such as undue bone–implant pressure, compromised bone
stability or increased bone removal. The most popular competing solutions are rescue
screws (a screw with a larger diameter)[11 ] and locking plate systems (systems that lock the screw and plate together).[7 ]
[15 ] However, current solutions have not adequately addressed screw loosening by these
hardware advancements.[7 ]
[13 ]
[14 ]
Accordingly, a novel rescue screw technology has been proposed that directly engineers
the bone–screw interface ([Fig. 1 ]). Succinctly, a unique bio-textile was fabricated into a braided sleeve; placed
around the screw increasing the surface area of contact between the screw and the
bone, thus enhancing screw engagement to prevent loosening.[16 ] The screw retention technology (SRT) studied here uses a cylindrical braided device
composed of polyethylene terephthalate mono filaments; a member of the polyester family
with no additives, it is not bioabsorbable.[16 ] Polyethylene terephthalate has previously demonstrated biocompatibility in many
clinical applications, including cardiovascular grafts,[17 ] plastic surgery application,[18 ]
[19 ] artificial ligaments[20 ]
[21 ] and bone augmentation.[22 ]
Fig. 1 Example image of a novel rescue screw retention technology designed to directly engineer
the bone–screw interface. The implant comprises a unique bio-textile (i.e. a braided
sleeve is placed around the screw).
In this application, the bio-textile interface provides a compliant layer between
the screw and bone; mechanical loads are distributed to reduce the pressure-induced
bone resorption that frequently occurs at the screw–bone interface.
The objective of this study was to investigate the ex vivo biomechanics and histological composition of ovine metatarsal fracture model treated
with a fixation plate and the SRT. The post-implantation ex vivo breakout and pullout strength biomechanics were determined. In addition, the bone
ingrowth adjacent to screws and callus healing was evaluated via histomorphometry.
Materials and Methods
This investigation was approved by the Institutional Animal Care and Use Committee
(IACUC no. 16-6379). This study used 24 skeletally mature sheep that underwent a unilateral
ostectomy and subsequent hardware implantation on their right metatarsus. Three time
points at 3, 6 and 12 weeks postoperatively were used; six sheep were sacrificed at
each of the three time points.
Sheep metatarsal bones underwent a 3 mm mid-diaphyseal transverse ostectomy that was
stabilized with a 9-hole, 3.5 mm LC-DCP plate (DePuy Synthes; West Chester, Pennsylvania,
United States) using seven proximally placed bicortical 3.5 mm cortical screws (DePuy
Synthes; West Chester, Pennsylvania, United States) and two distally placed unicortical
4.0 mm cancellous screws (DePuy Synthes). To simulate a scenario requiring rescue
screws, 3.5 mm pilot holes were drilled for the seven proximal bicortical screws and
4.0 mm pilot holes were created for each of the distal screws. This surgical model
was utilized to induce the worst possible case of screw–bone engagement (e.g. no screw–bone
engagement without SRT augmentation). The SRT device (OGmend Implant System; Woven
Orthopedic Technologies, LLC Manchester, Connecticut, United States) was slid over
the outer diameter of all screws.[16 ] The length of the SRT was matched to the length of the screw body and placed into
the pilot hole using a stylus, leaving ≤ 1 mm portion of the device exposed. The exposed
portion of the SRT implant was monitored to ensure the device did not migrate.
Terminal in vivo insertion torque (N-m) was measured during surgery using a torque sensing screwdriver
(TAT300; Futek, Inc., Irvine, California, United States). Following healing, euthanasia
and gross dissection, freshly harvested metatarsal samples were prepared; extraneous
soft-tissue was removed, taking great care not to damage the screw insertion, SRT
implant, callus or ostectomy.
Following gross dissection, the bone samples were photographed and subject to biplanar
digital radiography ([Fig. 2A ] and [2B ]). Screw breakout torque (defined as the initial break-out moment [N-mm] required
to loosen the screw) was also determined for n = 3 of 9 screws per sample by the digital torque sensing screwdriver.
Fig. 2 Example images of metatarsals in the study, including (A ) radiograph at time 0, (B ) photograph following dissection, (C ) radiograph prior to screw pullout showing cut plate and (D ) radiograph following screw pullout showing remaining screws for histology.
Destructive screw pullout force was also determined for n = 3 of 9 screws per sample. Fracture fixation plates were cut using a rotary cutoff
wheel to isolate each screw ([Fig. 2C ]). Care was taken to ensure that the isolation process did not detrimentally affect
the adjacent screws; samples were irrigated with saline to minimize thermal effects.
Metatarsi were then rigidly mounted into a testing system (Mini Bionix 858, MTS System,
Eden Prairie, Minnesota, United States). The long axis of the screw was aligned collinearly
with the actuator to ensure a normal vector pullout direction. Individual screw–plate
constructs were quasi-statically withdrawn at a rate of 1 mm/s. Force (N) and displacement
(mm) data were collected at 150 Hz. Construct stiffness (N/mm) and ultimate failure
load (N) were calculated for samples allocated to pullout testing. A total of 18 screws
were investigated each for breakout and pullout at each time point.
Histological analyses were conducted on the ostectomy fracture to demonstrate the
quality of healing at the bone defect site, contributing six tissue samples per time
point. Osteotomy site samples were processed using standard decalcified paraffin techniques
and stained with haematoxylin and eosin.
The remaining three screws in each metatarsal ([Fig. 2D ]) were used for non-decalcified hard tissue histology, yielding a total of 18 screws
for histological evaluation at each time point. Samples were processed using standard
non-decalcified techniques.[16 ]
[23 ]
[24 ] Sections were taken along the long axis of the screw to display the implant and
surrounding bone. Sections were stained with Sanderson's Rapid Bone, and then counterstained
using Van Gieson's solution.
Images were acquired for the entire section using a microscope (AG Heinze; Lake Forest,
California, United States) and digital camera (Diagnostic Instruments; Sterling, Heights,
Michigan, United States). Image Pro software (Media Cybernetics, Silver Spring, Maryland,
United States) was used for histomorphometric measurements.
The ostectomy region of interest (ROI) was set as the area from the proximal surgical
osteotomy cut to the distal cut. The screw ROI was set as the length of screw within
the bone with an area extending 300 µm towards the centreline of the screw and 300 µm
into the native bone. The histomorphometric parameters measured within each ROI were
percent bone area within the ROI (%), percent fibrous/soft tissue within the ROI (%),
percent implant (as applicable for screw and SRT device) area within the ROI (%) and
percent void area within the ROI (%).
Histology sections were also evaluated by a certified pathologist to document the
cellular responses for each of the samples. The pathologist was blinded to the treatment
group. The sections were qualitatively analysed according to cell type (i.e. polymorphonuclear,
lymphocytes, plasma, macrophages, giant and osteoblastic cells) and implant responses
(i.e. signs of bone remodelling, implant degradation and neovascularization).
Significance was determined using a standard one-way analysis of variance (ANOVA)
test, where p -values less than 0.05 were considered to be significant (SigmaStat; Systat Software
Inc., San Jose, California, United States). A post-hoc Student–Newman–Keuls multiple comparison analysis was performed to determine statistically
relevant p -values. Significant differences are designated with similar letters. The statistical
power for any comparison was above 0.80.
Results
Clinical assessment, as observed by two board certified veterinary surgeons (J.T.E
and R.H.P.), of screw insertion in the oversized holes indicated that surgical screw
insertion with SRT augmentation felt surgically tight and clinically acceptable during
in vivo implantation. All animals survived to term and gross necropsy yielded no adverse
findings. Post-sacrifice radiographs indicated normal osteotomy healing, no screw
backout or plate migration. All biomechanical and histological tests were run to completion
and no experimental issues were noted.
For cortical screw trajectories, the terminal insertion torque (mean ± one standard
deviation) with the SRT augmentation was 0.34 ± 0.10 N-m, 0.43 ± 0.15 N-m and 0.37 ± 0.12
N-m for 3, 6 and 12-week groups respectively. Similarly, the terminal insertion torque
for cancellous screws was 0.45 ± 0.20 N-m, 0.61 ± 0.27 N-m and 0.36 ± 0.17 N-m for
3, 6 and 12-week groups respectively. No significant difference in insertion torque
was found between groups for either the cancellous or cortical screws.
Three-, 6- and 12-week group cortical screw breakouts exhibited mean torques of 0.07 ± 0.03
N-m, 0.14 ± 0.08 N-m and 0.15 ± 0.09 N-m, respectively. Despite no significant difference
between the 6- and 12-week groups, both 6- and 12-week groups had significantly larger
breakout torque magnitudes as compared with the 3-week group (p < 0.01).
Screw pullout force (N) and stiffness (N/mm) data are presented ([Fig. 3 ]). The mode of failure was consistent across sacrifice time points with screws failing
under straight axial displacement with mild/moderate bone avulsion. No micro-motion
at the screw–plate interface was observed.
Fig. 3 Screw pullout data of metatarsals treated with screw retention technology after 3,
6 and 12 weeks, showing (A ) cortical screw pullout force (A, B, C: p < 0.01), (B ) cancellous screw pullout force (A, B: p = 0.02), (C ) cortical screw pullout stiffness (A, B: p < 0.01), and (D ) cancellous screw pullout stiffness (A: p = 0.01).
Cortical and cancellous pullout forces for screws with cortical trajectories are shown
([Fig. 3A ] and [3B ], respectively). Cortical screw pullout forces were significantly different, with
pullout forces significantly increasing between all three sacrifice time points (all
p < 0.01). Cancellous pullout forces also demonstrated significant increases in magnitude
as a function of healing time (p = 0.02), with the lone exception that there was not a significant increase between
the 6- and 12-week time points (p = 0.06).
The results of histomorphometric analyses for screw ROI and ostectomy ROI are shown
in [Tables 1 ] and [2 ], respectively). No statistical differences were calculated for total ROI areas across
sacrifice time points (p -values of 0.93, 0.84 and 0.73 for the cortical screw, cancellous screw and osteotomy
ROI respectively).
Table 1
Histomorphometric data (mean ± standard deviation) for cortical and cancellous screw
ROI
Constituents of interest
Cortical screws
Cancellous screws
3 Wk
6 Wk
12 Wk
3 Wk
6 Wk
12 Wk
% Bone
16.60 ± 3.70A
17.90 ± 4.90B
21.5 ± 5.40A,B
10.10 ± 4.30C
14.80 ± 5.30D
17.90 ± 6.60C,D
% Soft tissue
8.06 ± 4.13E
8.88 ± 6.40F
5.42 ± 3.32E,F
12.8 ± 4.70G,H
6.42 ± 2.80G
7.74 ± 4.67H
% Implant
3.35 ± 1.73
2.99 ± 1.53
3.04 ± 1.95
2.74 ± 1.14
2.18 ± 1.10
3.07 ± 1.35
% Screw
57.6 ± 7.20
53.7 ± 11.40
58.90 ± 5.90
47.10 ± 7.60
42.10 ± 14.90
43.40 ± 9.50
% Void space
14.40 ± 5.60I
16.60 ± 6.90J
11.20 ± 4.50I,J
27.20 ± 5.50
34.50 ± 17.70
27.80 ± 9.60
Abbreviation: ROI, region of interest.
Note: Significant differences are indicated by like letters (A, B, D, E, G, H, I,
J: p < 0.01; C: p < 0.04; F: p < 0.03).
Table 2
Histomorphometric data for osteotomy ROI
Constituents of interest
3 Wk
6 Wk
12 Wk
% Bone
43.20 ± 12.80
46.70 ± 14.80
53.40 ± 9.70
% Soft tissue
13.00 ± 8.00
16.20 ± 12.80
14.00 ± 9.40
% Void space
43.70 ± 10.70
38.40 ± 14.30
32.60 ± 10.90
Abbreviation: ROI, region of interest.
Note: No significant differences were found.
Histopathology showed that the SRT sleeves were embedded within reactive fibrosis
and associated with a rare population of lymphocytes and few macrophages. The SRT
sleeve was observed to be embedded within the bone in most cases ([Fig. 4 ]). Qualitative histopathology analysis indicated that there were no signs of abnormal
gross cellular reactions (inflammation or infection) at the bone–screw–SRT interfaces
or osteotomy sites. There were also no gross signs of device degradation or debris
indicating the device maintained its structural integrity throughout the study.
Fig. 4 (Top left and top middle) Example histologic images demonstrating the histomorphometric
region of interest (i.e. the screw–bone interface) at 10x magnification at 6 and 12
weeks post-implantation. (Bottom left and bottom middle) Example images highlighting
the implant, screw retention technology (SRT) device, bone and soft tissue at the
screw–bone interface at 100x magnification at 6 and 12 weeks post-implantation. (Right ) Example image (derived from an unrelated study; unpublished data) showing a typical
control screw–bone interface in the ovine metatarsal at 12 weeks post-implantation.
Discussion
Pullout and breakout biomechanical testing exhibited consistent improvements in screw
retention as a function of increased healing times. Histologically, increased bone
fraction surrounding the cortical and cancellous screws consistent with improved screw
retention was observed following healing. Increased bone fraction and decreased void
fraction, as well as radiographic changes, were identified at the ostectomy indicating
successful progression of osteotomy healing toward osseous union consistent with metatarsal
osteotomy healing previously observed by our group utilizing standard screw–bone engagement
models.[25 ]
[26 ]
[27 ]
[28 ]
Unfortunately, no previous studies in the literature were found on screw pullout and
breakout in ovine metatarsals for direct comparison. However, sheep metatarsi are
commonly used as an analogue for human tibia due to similarities in size and bone
mineral density between the two bones.[25 ]
[26 ]
[27 ]
Matityahu and colleagues[29 ] evaluated the pullout strength of 3.5-mm self-tapping screws in a standard drill
hole model using cadaveric diaphyseal tibiae. It was found that a single insertion,
four insertions and five insertions yielded pullout strengths of 1710 ± 550 N, 1030 ± 543 N
and 364 ± 209 N respectively. These findings are of a similar magnitude to the pullout
strengths of the present study and indicate that, after 12 weeks of healing, the SRT
provides similar screw retention to screws with one insertion and no over drilling.
In addition, our data indicated that after 3 weeks of healing, the SRT provides greater
screw retention than screws with five insertions and no overdrilling. Similarly, Oldakowska
and colleagues[30 ] evaluated the pullout strength of 4 mm self-tapping screws in a standard drill hole
model in cadaveric thoracic ovine bone. It was found that the screws exhibited a failure
force of 695.0 ± 82.4 N, and a stiffness of 618.5 ± 114.1 N/mm. Again, these findings
are of a similar magnitude to the pullout and stiffness data generated for this study,
which utilized an overdrilled hole model, following 6 and 12 weeks of healing. The
similarity between our data and these studies indicates that the SRT may allow relevant
healing at the screw–bone interface in an overdrilled (i.e. rescue) screw scenario
following healing.
Claes and colleagues[31 ] explored the tissue differentiation of ovine metatarsal ostectomies (2.1 mm) following
9 weeks of healing subject to the established method of external fixation. Despite
discrepancies in healing times and fracture size, all values in the present study
were within one standard deviation of all reported cortical and medullary tissue distributions
for bone, soft tissue and connective tissue/void space. Indeed, even the 3-week time
point in the present study observed similar ostectomy tissue differentiation to a
smaller fracture after a longer healing duration. Similarly, Augat and colleagues[32 ] also investigated external fixation of 2.0 mm ovine metatarsal ostectomies. Following
9 weeks of healing, histological data at the ostectomy yielded similar composition
of bone, soft tissue and connective tissue/void space to the present study. Histomorphometric
data indicated the percent bone increased, the percent soft tissue did not increase
and the percent implant remained constant; these trends likely resulted from normal
bone remodelling, lack of excessive fibrotic reaction to the SRT and no degradation
of the SRT. This increase in bone likely resulted in the observed increases in mechanical
integrity. The congruence of the tissue differentiation with the discussed studies
indicates that effective fracture healing was induced in the treated group of this
study. Therefore, it appears that the SRT was essential at promoting the appropriate
mechanical integrity to allow for fracture healing.
When complications occur, either intraoperatively or in revision procedures, surgeons
must remove additional bone stock to replace the loose screws, thus limiting the ability
of the surgeon to generate the necessary stability and reduction for fracture fixation.
When orthopaedic screws require revision, surgeons use a variety of ad-hoc techniques including the use of larger and/or longer screws (i.e. rescue screw),
inserting screws in a different trajectory/pilot hole, use of additional plates or
augmenting the failed hole with bone void fillers or polymethyl methacrylate (i.e.
bone cements) or with the Matchstick method.[33 ] Unfortunately, the use of a strip ('matchstick') of bone graft to act as a shim
leads to asymmetric hoop stresses with force concentration at the strip of bone graft
and increases the risk of screw hole wall fracture particularly with compromised (osteoporotic)
bone quality. Cements also have several drawbacks. A major problem is the difficulty
of precise placement and the prevention of inadvertent migration of the semi-liquid
cement which could cause problems (i.e. mechanical impingement). In addition, the
most commonly used orthopaedic cement, polymethyl methacrylate, in the process of
in situ polymerization can cause local or systemic toxicity as in the case of monomer release.
Also, there is the generation of high local temperatures as the cement exothermically
polymerizes, with the potential for thermal injury to local structures. The main problem
with repositioning approaches is the creation of stress risers with the empty screw
holes weakening the underlying osseous structure; local anatomic considerations may
also prevent plate repositioning. Therefore, there still exists a clinical need for
a device which can prevent screw complications and revisions, and, if revision is
required, provides a more robust stabilization augmentation in clinical surgery. The
SRT device studied here does not suffer from these issues. The structure of the device
maximizes screw bone engagement in a uniform circumferential fashion avoiding stress
concentration. The SRT has no issues with either exothermic curing, or local toxicity
as the material of the SRT has well-documented history of bio-compatibility. With
the SRT, there is no need to leave a screw hole empty eliminating subsequent stress
riser creation and obviating the need to reposition an orthopaedic plate or device.
The design of this study excluded an in vivo negative control group (i.e. overdrilled pilot hole with a standard screw). Subjectively,
all screws in a negative control group were considered clinically unacceptable and
unsafe for fracture repair in an in vivo setting by two board-certified veterinary surgeons. This was further validated by
time-zero ramp to failure testing on cadaveric ovine metatarsal samples (data not
shown); the measured failure load for a proposed negative control group was determined
to be less than the estimated load on the treated limb at any time during in vivo healing. Accordingly, application of a negative control group in an in vivo animal model would have been ill-advised, in-humane and against the spirit of IACUC
guidelines. It was determined that any negative control group samples would suffer
catastrophic failure upon standing and full-weight-bearing immediately following recovery
from surgery. A positive control (i.e. a screw placed in the surgically standard hole
or the use of some form of rescue screw) was deemed to be the most appropriate control
for this model. However, as this study was an initial attempt to prove the efficacy
of the SRT device a positive control arm of the study was not implemented. By comparing
our results to data from previous studies,[29 ]
[30 ] that used the current standards of screw insertion, it appears that augmentation
of an overdrilled hole with the SRT leads to similar levels of acute biomechanical
stability following healing. This assertion is further strengthened by the fact that
all fractures generated within this study had typical healing pathways leading towards
clinically acceptable osseous union, which would be unlikely if the implanted hardware
was not adequality stabilizing the fracture. A recent study examining the effectiveness
of the SRT in an ovine spine model, in which a positive and negative control where
possible, demonstrated that the SRT device does improve screw–bone purchase as compared
with a negative control.[16 ] However, while the literature is replete with studies that have translated the general
results of sheep orthopaedic models to human applications,[26 ]
[27 ]
[34 ]
[35 ] one should take caution when prescribing the absolute values of ovine-derived data
to that of the human condition.
In conclusion, the novel SRT investigated in this studied showed improved screw retention
in an in vivo ovine metatarsal model with oversized holes as healing time progressed, and that
the biomechanical stability imparted by the SRT device was of the same order observed
for standard screws implanted acutely. The biomechanical and histological results
of this study demonstrate the SRT as an effective method for improved screw retention
for mitigation of clinical screw failure in situations that might otherwise have clinically
unsatisfactory fixation.
