Biomechanics and Modeling in Mechanobiology (2020) 19:693–699

https://doi.org/10.1007/s10237-019-01240-3

ORIGINAL PAPER

Stress analysis in a bone fracture fixed with topology‑optimised plates

Abdulsalam Abdulaziz Al‑Tamimi1,2 · Carlos Quental3 · Joao Folgado3 · Chris Peach2,4 · Paulo Bartolo2

Received: 10 July 2019 / Accepted: 9 October 2019 / Published online: 24 October 2019
© The Author(s) 2019

Abstract
The design of commercially available fixation plates and the materials used for their fabrication lead to the plates being
stiffer than bone. Consequently, commercial plates are prone to induce bone stress shielding. In this study, three-dimensional
fixation plates are designed using topology optimisation aiming to reduce the risk of bone stress shielding. Fixation plate
designs were optimised by minimising the strain energy for three levels of volume reduction (i.e. 25%, 45% and 75%). To
evaluate stress shielding, changes in bone stress due to the different fixation plate designs were determined on the fracture
plane of an idealised shaft of a long bone under a four-point bending load considering the effect of a patient walking with
crutches of a transverse fractured tibia. Topology optimisation is a viable approach to design less stiff plates with adequate
mechanical strength considering high volume reductions, which consequently increased the stress transferred to the bone
fracture plane minimising bone stress shielding.

Keywords Fixation plates · Stress analysis · Stress shielding · Topology optimisation

1 Introduction 2008; Quental et al. 2014). This means that the load dis-
tribution in the bone-plate interface during healing will be
Stress shielding is an important phenomenon that must uneven, mainly supported by the bone plate and screws.
be considered during design optimisation of fracture fixa- This will shield the bone from the stress stimulus required
tion plates to minimise the risk of bone resorption and to provide adequate bone healing and eventually cause bone
plate failure (Prasad et al. 2017). It is the result of the stiff- resorption and implant loosening, in a phenomenon known
ness mismatch between the most commonly used metallic as “stress shielding” (Ridzwan et al. 2007; Prasad et al.
fracture fixation plates and bones (e.g. Young’s modulus 2017).
of Ti–6Al–4V is around 120 GPa and cortical bone is Stress shielding in bone is a common problem induced
15–25 GPa), which strongly determines the bone remodel- by mild- to high-load-bearing medical implants and can be
ling process whereby, according to Wolff’s law, bone adapts reduced by redesigning the medical implant (Ramakrishna
to the forces acting upon it (Ridzwan et al. 2007; Elias et al. et al. 2004; Galbusera et al. 2009). The use of topology opti-
misation is gaining significant attention due to the ability to
automatically generate optimal redesigns for a given design,
Electronic supplementary material The online version of this
article (https​://doi.org/10.1007/s1023​7-019-01240​-3) contains considering different loading conditions and volume reduc-
supplementary material, which is available to authorized users. tion constraints. Several authors demonstrated the feasibility
of topology optimisation for the redesign of orthopaedic medi-
* Paulo Bartolo cal implants to minimise stress shielding such as femur hip
paulojorge.dasilvabartolo@manchester.ac.uk
joints (Ridzwan et al. 2006; Fraldi et al. 2010; Saravana and
1
Industrial Engineering Department, College of Engineering, George 2017), spine (Chuah et al. 2010) and pelvic prostheses
Kind Saud University, Riyadh, Saudi Arabia (Iqbal et al. 2019). In these cases, results showed improved
2
School of Mechanical, Aerospace and Civil Engineering, The load transfer to the bone in the case of optimised implants.
University of Manchester, Manchester, UK Similarly, Liu et al. (2017) used topology optimisation to
3
IDMEC, Instituto Superior Técnico, Universidade de Lisboa, design mandible fixation plates with adequate biomechanical
Lisbon, Portugal performance. The optimised designs obtained through topol-
4
Manchester University NHS Foundation Trust, Manchester, ogy optimisation present complex internal/porous structures
UK

13
Vol.:(0123456789)
694 A. A. Al‑Tamimi et al.

difficult to produce using conventional manufacturing tech-

⎧ ∑ 𝜌 v ≤ V ∗,
N
nologies (Iqbal et al. 2019). Additive technology is the ideal
technique to produce these plates, not only due to the ability to ⎪ e=1 e e
⎪� N �
produce very complex shapes, but also due to the fact that the subject to ⎨ ∑ p (1)
use of additive manufacturing allows to reduce material waste, ⎪ e=1 𝜌e 𝐊e 𝐮 = 𝐟,
part fabrication without the use of complex tooling, being the ⎪
⎩ 0 < 𝜌0 ≤ 𝜌e ≤ 1,
ideal technology for mass personalisation (Parthasarathy 2015;
Murr 2016). For metal fixation plates, the ideal additive manu-
facturing technologies are selective laser melting (SLM) and where C is the compliance, p is a penalisation factor (p = 3),
electron beam melting (EBM) (Calignano et al. 2019; Yuan u is the displacement vector, f is the force vector, ρe is the
et al. 2019). element density, ρo is the initial density, V* is the volume
A valid concern when reducing stiffness of implants is fraction, ve is the volume of each element, and Ke is the
whether the change in biomechanical characteristics has a element stiffness matrix. The topology optimisation prob-
negative effect on the stability at the fracture site which might lem was run using the TOSCA module in Abaqus (Dassault
affect bone healing. In order to ensure appropriate stresses Systèmes, Waltham, MA, USA).
imposed on the bone, fixation plates’ stiffness should be opti- The mechanical behaviour of both initial and optimised
mised whilst maintaining plate stability during the healing designs was investigated through finite element analyses,
process. assuming elastic behaviour and homogeneous and iso-
Therefore, this paper, as the authors’ best of knowledge, is tropic plates. Numerical simulation was used to determine
the first study investigating the use of topology optimisation the equivalent stiffness in a four-point bending setting
to design fixation plates that minimise stress shielding and according to the British standards (BS 3531-23.1:1991
promote load transfer to the bone fracture plane, thus stimu- ISO 9585:1990). In this case, the equivalent bending stiff-
lating bone remodelling. Two different fixation plates were ness is determined according to the following equation:
considered (four- and eight-screw holes) and were optimised ( 2 )
4h + 12ha + a2 ⋅ K ⋅ h
for different loading conditions (bending, compression, torsion EB = (N m2 ) (2)
and a combined load) and three volume reductions (25%, 45% 24
and 75%). The optimised plates were evaluated using finite ele- where h is the distance between the force points, a is the
ment analyses considering a tibia-like bone shape model under span between the force and support points, and K is the stiff-
a bending loading condition to study the induced stresses on ness calculated as follows:
the defined bone fracture plane.
RF
K= (N/m) (3)
D
2 Modelling and simulation where RF is the average reaction force in the z axis (along
with the fixation plate thickness) at the constraint points and
2.1 Optimisation and mechanical evaluation D is the maximum displacement.
Changes in the equivalent stiffness between the opti-
The Solid Isotropic Microstructure with Penalisation (SIMP) mised fixation plates and the initial designs were calcu-
method was applied to redesign two different fixation plates lated using the following equation:
with four- and eight-screw holes. Initial designs were cre-
ated in Solidworks (Dassault Systèmes, Waltham, MA, USA) Δequivalent stiffness (%)
( )
considering generic locking compression fixation systems Plate stiffnessOptimised − Plate stiffnessInitial (4)
for treating long bones for midshaft fractures with a length =
Plate stiffnessInitial
of 180 mm, width of 14 mm and thickness of 5 mm. Three
volume reductions (25%, 45% and 75%), and different loading
2.2 Stress analysis of a bone model
conditions (bending, compression, torsion and a combination
of all these loads), were considered as shown in Fig. 1. A fro-
In order to determine the stresses in the fracture plane of
zen region was considered on the screw hole region to keep
a bone, which provides an indication of the stress shield-
their shape. Simulations were performed considering quadratic
ing effect of the plates, a bone-plate construct (i.e. the
hexahedral meshes with around 50,000 elements.
assembly of the fracture plate to the bone with screws) was
Mathematically, the SIMP formulation can be described as
considered. For simplicity, no fracture gap was imposed,
follows (Bendsøe and Sigmund, 2004):
and the plate was assumed to fixate a transverse fractured
( ) tibia bone. Only the cortical bone region was considered,
min C 𝜌e = 𝐟 T ⋅ 𝐮
𝜌e

13
Stress analysis in a bone fracture fixed with topology‑optimised plates 695

Fig. 1  Load and boundary

conditions considered for the
fixation plate optimisation
of a four-point bending load,
b uniaxial compression, c
torsional and d combination of
the bending, compression and
torsion loads

13
696 A. A. Al‑Tamimi et al.

and a hollow cylinder region with an external diameter of fixation plates were used to determine their mechanical
24 mm and an internal diameter of 12 mm was assumed strength and, consequently, to investigate their stability.
for simulation purposes.
All 3D geometric parts, i.e. cortical bone, screws and ini-
tial fixation plates, were modelled in Solidworks (Dassault 3 Results
Systèmes, Waltham, MA, USA). Each screw has a 5-mm-
diameter head and a main body with 3.5 mm of diameter The changes in the equivalent stiffness between the opti-
and 34 mm of total length. Both fracture plates and screws mised and initial designs are shown in Table 1. Optimised
were assumed to be made of Ti–6Al–4V. For the cortical designs obtained through topology optimisation are pre-
bone, a Young’s modulus of 18 GPa and a Poisson’s ratio sented as supplementary material (Figs. S1 and S2). The
of 0.3 were assumed (Santos et al. 2018). In order to avoid equivalent bending stiffness change increases as the vol-
high computational costs, only half of the bone-plate con- ume reduction increases. For the same volume reduction
struct was considered, as illustrated in Fig. 2. To simulate and plates with different number of holes, there is no clear
the Locking Compression Plate technique, the finite element
model considered the bone-plate construct with a gap of
Table 1  Change in the equivalent bending stiffness in compari-
0.5 mm between the bone and plate (i.e. no contact). The son with the initial values for four- and eight-hole plates (19.27 and
screw heads were securely locked to the plate and the screws 16.22 N m2, respectively)
tied to the bone. Quadratic hexahedral elements were con-
Plate Volume Hole numbers
sidered for the bone model region of interest (i.e. fracture reduction (%)
plane) and quadratic tetrahedron elements for the plates, Equivalent stiffness change (%)
screws and the bone region outside of the fracture plane. Four-hole plate Eight-hole plate
Two equally distributed moments of 20 Nm were applied
Bending 25 −3 −2
along the horizontal axis of the bone, simulating the moment
45 −5 −5
load happening on the tibia during the swing phase (i.e. 10%
75 − 20 − 61
of the body weight) in patients walking with crutches (Ram-
Compression 25 − 15 − 15
akrishna et al. 2004; Wehner et al. 2009; Kim et al. 2011).
45 − 17 − 28
To prevent rigid body motion, the extremity faces of the
75 − 49 − 71
bone were fully constrained.
Torsion 25 −3 − 31
In addition, since the mechanical strength of the topology-
45 −9 − 33
optimised fixation plates is important to assess their stabil-
75 − 49 − 64
ity during healing, a mechanical strength analysis was per-
Combined 25 −4 −4
formed based on the materials yield strength, considering the
45 − 57 −7
yield strength of the Ti–6Al–4V as ~ 860 MPa (Elias et al.
75 − 92 − 87
2008). The von Mises stresses on the topology-optimised

Fig. 2  Bone-plate construct

considered to determine stresses
at the fracture plane

13
Stress analysis in a bone fracture fixed with topology‑optimised plates 697

trend in terms of the equivalent stiffness change. Also, as

expected, plates optimised considering bending loading
conditions exhibit the least reduced equivalent stiffness
change. The highest decreases were observed for the 75%
volume reduction with combined load for the four-hole and
eight-hole.
The maximum von Mises stresses from the bone-plate
construct at the fracture plane, considering both the initial
and topology-optimised fixation plates, are presented in
Table 2. The stress distribution and magnitude at the bone
fracture plane are presented in Fig. 3, for the initial designs
and only for the 75% volume reduction optimised plates,
which is the case causing the maximum von Mises stresses
on the bone at the fracture plane. Overall, the stresses in the
bone increase when topology-optimised plates are used. The
most substantial increase in stresses at the fracture plane was
observed for the combined loading conditions and 75% vol-
ume reduction plates. In comparison with the initial designs,
the maximum stresses at the fracture plane increased by 31%
for the four-hole plate and 37% for the eight-hole plate in
the combined case with a 75% of volume reduction. The
stress distribution shows that less stiff plates produce higher
compressive stresses (in the plate-bone interface due to the
bending load) at the fracture plane. Moreover, the neutral
axis with the less stiff plates becomes closer to that of the
bone’s when compared with high stiff plates. In terms of
plate mechanical stability (i.e. mechanical strength), the
stresses among the least stiff fixation plates (75% of volume
reduction) are shown in Fig. 4. The maximum stresses occur
on both four-hole and eight-hole designs with a combined
load and 75% of volume reduction. Minimum values of the

Table 2  Maximum von Mises stresses on the bone at the fracture

plane for all considered designs
Plate Volume Hole numbers Fig. 3  von Mises stresses at the bone fracture plane resulted from the
reduction initial designs and all of the 75% volume reduction optimised plates
(%) von Mises stress (MPa)
Four-hole plate Eight-hole plate
von Mises stresses for all plates occur for compression loads
Initial designs N/A 17.59 15.37
and 75% of volume reduction.
Bending 25 17.95 15.38
45 19.93 15.93
75 21.41 18.15
4 Discussion and conclusions
Compression 25 17.94 15.46
45 19.89 15.58
Results show that through topology optimisation it is pos-
75 20.79 19.91
sible to design less stiff fixation plates, resulting in higher
Torsion 25 17.77 15.73
loads being transferred to the bone fracture plane, whilst
45 19.92 15.81
maintaining the plate’s ability to withstand stresses. By con-
75 20.03 18.25
sidering a maximum of 75% of volume reduction for plates
Combined 25 17.61 15.40
containing different screw holes, it is possible to increase
45 17.88 15.75
the load transfer to up to 37% in comparison with the initial
75 22.96 21.07
plates. This reduces stress shielding, and likely bone loss, to

13
698 A. A. Al‑Tamimi et al.

Fig. 4  Stress distribution on the

optimised four- and eight-screw
hole plates with 75% volume
reduction and different load-
ing conditions a bending, b
compression, c torsion and d a
combined load

promoting secondary healing, that promotes callus forma- load transfer of 29% and 240 MPa of von Mises stresses
tion and bone formation (Woo et al. 1977; Goodship and were observed for compression load optimised plates.
Kenwright 1985; Claes et al. 1997). As shown, topology optimisation permits the design of
Maximum von Mises stresses were observed in plates less stiff and lightweight fixation plates, reducing the stress
optimised for combined loading conditions and 75% of vol- shielding effect, promoting load transfer to the bone and thus
ume reduction. This can be explained by stress concentra- contributing to bone remodelling. However, further analysis
tions induced by the plate design presenting thin features. is still required, considering for example a fracture gap and
However, as observed, the highest stresses are still 50% measuring the gap strains to correlate the resulted strains
lower than the yield strength of the material, guaranteeing (i.e. relative or absolute stability) with the healing process
the plate mechanical stability. (i.e. secondary or primary healing). Furthermore, screw
For the four-hole plates, the best performance was threads were not considered in the simulation and their role
observed for plates with 75% of volume reduction and opti- on load transfer must be also considered. Although no criti-
mised for bending loading conditions, which enable 22% cal failure was observed in the optimised designs, thin and
of load transfer to the bone, presenting also low von Mises sharp features were yet observed. Further post-processing
stresses (221 MPa). For the eight-hole plates, a maximum design steps must be considered to address these features.

13
Stress analysis in a bone fracture fixed with topology‑optimised plates 699

Additional constraints must also be considered to allow fur- Kim S, Chang S, Son D (2011) Finite element analysis of the effect of
ther fabrication (e.g. thickness of internal features which bending stiffness and contact condition of composite bone plates
with simple rectangular cross-section on the bio-mechanical
might be difficult to produce using additive manufacturing). behaviour of fractured long bones. Compos Part B Eng 42:1731–
1738. https​://doi.org/10.1016/j.compo​sites​b.2011.03.001
Acknowledgements Authors are grateful for the support provided by Liu Y, Fan Y, Jiang X, Baur D (2017) A customized fixation plate with
the Saudi Arabian government, the Manchester University Foundation novel structure designed by topological optimization for mandibu-
NHS Trust, Royal College of Surgeons of England, and the Portu- lar angle fracture based on finite element analysis. BioMed Eng
guese Foundation for Science and Technology, through IDMEC, under OnLine 16:1–17. https​://doi.org/10.1186/s1293​8-017-0422-z
LAETA, project UID/EMS/50022/2019. Murr LE (2016) Frontiers of 3D printing/additive manufacturing: from
human organs to aircraft fabrication. J Mater Sci Technol 32:987–
Compliance with ethical standards 995. https​://doi.org/10.1016/j.jmst.2016.08.011
Parthasarathy J (2015) Additive manufacturing of medical devices. In:
Srivatsan TS, Sudarshan TS (eds) Additive manufacturing inno-
Conflict of interest The authors declare that they have no conflict of vations, advances, and applications. CRC Press, Boca Raton, pp
interest. 369–388
Prasad K, Bazaka O, Chua M, Rochford M, Fedrick L, Spoor J, Symes
Open Access This article is distributed under the terms of the Crea- R, Tieppo M, Collins C, Cao A, Markwell D, Ostrikov K, Bazaka
tive Commons Attribution 4.0 International License (http://creat​iveco​ K (2017) Metallic biomaterials: current challenges and oppor-
mmons​.org/licen​ses/by/4.0/), which permits unrestricted use, distribu- tunities. Materials 10:884. https​://doi.org/10.3390/ma100​80884​
tion, and reproduction in any medium, provided you give appropriate Quental C, Fernandes PR, Monteiro J, Folgado J (2014) Bone remod-
credit to the original author(s) and the source, provide a link to the elling of the scapula after a total shoulder arthroplasty. Biomech
Creative Commons license, and indicate if changes were made. Model Mechanobiol 13:827–838. https​://doi.org/10.1007/s1023​
7-013-0537-5
Ramakrishna K, Sridhar I, Sivashakner S, Khong K, Ghista D (2004)
Design of fracture fixation plate for necessary and sufficient bone
References stress shielding. JSME Int J Ser C 47:1086–1094. https​://doi.
org/10.1299/jsmec​.47.1086
Bendsøe M, Sigmund O (2004) Topology optimization: theory, meth- Ridzwan M, Shuib S, Hassan A, Shokri A, Ibrahim M (2006) Optimiza-
ods, and applications. Springer, Berlin tion in implant topology to reduce stress shielding problem. J Appl
Calignano F, Galati M, Iuliano L, Minetola P (2019) Design of addi- Sci 6:2768–2773. https​://doi.org/10.3923/jas.2006.2768.2773
tively manufactured structures for biomedical applications: a Ridzwan M, Shuib S, Hassan A, Shokri A, Ibrahim M (2007) Prob-
review of the additive manufacturing processes applied to the lem of stress shielding and improvement to the hip implant
biomedical sector. J Healthc Eng 2019:9748212. https​://doi. designs: a review. J Med Sci 7:460–467. https​://doi.org/10.3923/
org/10.1155/2019/97482​12 jms.2007.460.467
Chuah HG, Rahim IA, Yusof MI (2010) Topology optimisation of Santos B, Quental C, Folgado J, Sarmento M, Monteiro J (2018)
spinal interbody cage for reducing stress shielding effect. Com- Bone remodelling of the humerus after a resurfacing and a stem-
put Methods Biomech Biomed Eng 13:319–326. https​://doi. less shoulder arthroplasty. Clin Biomech 59:78–84. https​://doi.
org/10.1080/10255​84090​32081​89 org/10.1016/j.clinb​iomec​h.2018.09.009
Claes L, Augat P, Suger G, Wilke H (1997) Influence of size and sta- Saravana KG, George S (2017) Optimization of custom cementless
bility of the osteotomy gap on the success of fracture healing. J stem using finite element analysis and elastic modulus distribu-
Orthop Res 15:577–584. https​://doi.org/10.1002/jor.11001​50414​ tion for reducing stress-shielding effect. Proc Inst Mech Eng Part
Elias C, Lima J, Valiev R, Meyers M (2008) Biomedical applications H J Eng Med 231:149–159. https​://doi.org/10.1177/09544​11916​
of titanium and its alloys. JOM 60:46–49. https:​ //doi.org/10.1007/ 68612​5
s1183​7-008-0031-1 Wehner T, Claes L, Simon U (2009) Internal loads in the human tibia
Fraldi M, Esposito L, Perrella G, Cutolo A, Cowin S (2010) Topologi- during gait. Clin Biomech 24:299–302. https​://doi.org/10.1016/j.
cal optimization in hip prosthesis design. Biomech Model Mecha- clinb​iomec​h.2008.12.007
nobiol 9:389–402. https​://doi.org/10.1007/s1023​7-009-0183-0 Woo S, Simon B, Akeson W, McCarty M (1977) An interdiscipli-
Galbusera F, Bertolazzi L, Balossino R, Dubini G (2009) Combined nary approach to evaluate the effect of internal fixation plate
computational study of mechanical behaviour and drug deliv- on long bone remodeling. J Biomech 10:87–95. https​://doi.
ery from a porous, hydroxyapatite-based bone graft. Biomech org/10.1016/0021-9290(77)90072​-0
Model Mechanobiol 8:209–216. https​://doi.org/10.1007/s1023​ Yuan L, Ding S, Wen C (2019) Additive manufacturing technol-
7-008-0132-3 ogy for porous metal implant applications and triple minimal
Goodship A, Kenwright J (1985) The influence of induced micro- surface structures: a review. Bioact Mater 4:56–70. https​://doi.
movement upon the healing of experimental tibial fractures. J org/10.1016/j.bioac​tmat.2018.12.003
Bone Joint Surg Br 67-B:650–655. https​://doi.org/10.1302/0301-
620x.67b4.40308​69 Publisher’s Note Springer Nature remains neutral with regard to
Iqbal T, Wang L, Li D, Dong E, Fan H, Fu J, Hu C (2019) A general jurisdictional claims in published maps and institutional affiliations.
multi-objective topology optimization methodology developed for
customized design of pelvic prostheses. Med Eng Phys 69:8–16.
https​://doi.org/10.1016/j.meden​gphy.2019.06.008

13