Journal of Biomechanics 31 (1998) 247252

A method to determine the 3-D stiness of fracture xation devices

and its application to predict inter-fragmentary movement
Georg N. Duda*, Helmut Kirchner, Hans-Joachim Wilke, Lutz Claes
Department of Unfallchirurgische Forschung und Biomechanik, University of Ulm, Ulm, Germany
Received in nal form 1 October 1997
Abstract
Inter-fragmentary movement considerably inuences the fracture healing process. Large shear movement delays while moderate
axial movement stimulates the healing process. To be able to control the mechanical situation at a fracture site and to achieve optimal
bony healing it is essential to understand the relationship between inter-fragmentary movement, bony loading and xation stiness.
A 6;6 stiness matrix is introduced which completely describes the linear relationship between the 6 inter-fragmentary movements
and the resulting bony loading (3 forces and 3 moments). Further, it is illustrated that even in relatively sti external xateur
constructs simple axial loading of the bony fragments leads to complex inter-fragmentary movement. When the 3-D stiness
description is multiplied by the load state in sheep tibiae, movements similar to those measured in vivo are calculated. The relation-
ship between axial compression and medio-lateral or dorso-ventral shear varies depending on the mounting plane of the external
xateur.
The authors conclude that a single value is not sucient to describe the mechanical relationship between inter-fragmentary
movement and bony loading. Only a complete description of xation stiness allows prediction of inter-fragmentary movement and
dierentiation between various congurations of xation devices and their potential for mechanically promoting bony healing.
1998 Elsevier Science Ltd. All rights reserved.
Keywords: Fracture healing; Inter-fragmentary movement; Fixation stiness; External xation
1. Introduction
Independent of the method of fracture xation, bone
healing is generally subjected to complex inter-fragmen-
tary movements. It is well accepted that this movement
inuences the fracture healing process both in its type
and rate of healing (Claes et al., 1995; Goodship and
Kenwright, 1985; Kenwright and Goodship, 1989).
Axial movement using xateur externe congurations
has been analyzed in various in vitro as well as in vivo
studies (Cunningham et al., 1989; Goodship et al., 1993;
Goodship et al., 1988; Homann et al., 1991; Kenwright
et al., 1991; Kristiansen et al., 1987; Lippert and Hirsch,
* Correspondence address: Forschungslabor der Unfall- und
Wiederherstellungschirurgie, Charite , Humboldt Universita t zu Berlin,
Augustenburger Platz 1, D-13353 Berlin, Germany. Tel.: #49 30 450
59079; fax: #49 30 450 59969; e-mail: duda@ukrv.de.
1974). It has been shown that a large stiness resulting in
small axial movements minimizes the risk of pseudar-
throses (Schenk et al., 1986; Stu rmer, 1988). However,
a certain amount of inter-fragmentary movement is ne-
cessary to achieve sucient mechanical stability in the
newly formed bone (Kenwright and Goodship, 1989;
Molster and Gjerdet, 1984; Molster et al., 1982).
Animal experiments have shown that an axial inter-
fragmentary movement within the range of 0.21.0 mm
seems to be optimal for fracture healing (Claes et al.,
1995; Goodship et al., 1988). However, it remains unclear
how fracture healing is related to the movement compo-
nents other than that in the axial direction. From in vivo
experiments and clinical experience the impact of axial
and shear movements on the healing callus has been
qualitatively described (Yamagishi and Yoshimura,
1955). Although a large number of xation devices in
various congurations are clinically used, the 3-D inter-
fragmentary movements actually occurring in vivo are
mainly unknown.
0021-9290/98/$19.00 1998 Elsevier Science Ltd. All rights reserved.
PI I S 0 0 2 1 - 9 2 9 0 ( 9 7 ) 0 0 1 1 5 - 2
Since axial and shear movements appear to inuence
the fracture healing processes dierently, it would be
benecial to know the 3-D inter-fragmentary movement
prior to mounting of the xation device (Homann et al.,
1991). Due to the highly asymmetric nature of xateur
externe devices an axially rigid construct may experience
large shear deformations under bending loads. Only
the complete description of the 3-D xation sti-
ness would allow the prediction of the full set of inter-
fragmentary movements occurring under complex
in vivo loading (Gardner et al., 1996). Since the bony
load state is primarily independent of the method of
fracture xation, knowledge of the constructs 3-D sti-
ness would allow dierentiation between those xation
divices that provide lesser or greater inter-fragmentary
movement. Fixation devices could then be congured
pre- or intra-operatively to meet the desired specica-
tions, e.g. axial movement between 0.2 and 1 mm, mini-
mized shear.
The goal of this study was to develop a method to
determine the 3-D stiness of fracture xation devices
and to predict inter-fragmentary movement as a function
of fracture location and xateur mounting.
2. Materials and methods
For determination of the 3-D xateur stiness, an
ASIF external xateur (double tube, steel rods, two
Schanz screws per fragment) was selected. The specic
dimensions are given in Fig. 1 and are identical to those
used previously in an in vivo recording of inter-fragmen-
tary movement in sheep (Stu rmer, 1988). The external
xateur was mounted to a pertinax rod (diameter 20 mm,
length 300 mm) which was then osteotomized at mid-
span between the inner Schanz screws, creating a 4-mm
fracture gap.
Prior to testing, a 3-D goniometer system (accuracy
0.1 mm, 0.1) was attached to the xateur conguration
using the proximal and distal Schanz screws (Fig. 2,
Wilke et al., 1994). A tight t of the Schanz screws in the
pertinax rod eliminated pin bone interface motion in this
in vitro experiment.
Mathematically, the loads between the proximal and
distal fracture fragments (three forces and three
moments) can be related to the inter-fragmentary
movements (three translations and three rotations) by
a stiness matrix. If the loads and gap movements are
written as column vectors each, the xation stiness is
described by a 6;6 stiness matrix with 36 unknown
stiness values [Eq. (1)]. These values are constant while
the load-displacement relationship is linear. To simplify
the relationship between the 3-D loading and 3-D inter-
fragmentary movement, the current analysis was per-
formed under the assumption that the entire system
behaved linear.
The diagonal values of the stiness matrix refer to the
ventral shear stiness, lateral shear stiness, compression
stiness bending stiness around an axis perpendicular,
bending stiness around an axis parallel and torsional
stiness.
To determine the 36 unknown values of the stiness
matrix six mechanically independent load cases (consist-
ing of 3 forces and 3 moments each) and correlating
inter-fragmentary movements (3 translations and 3 rota-
tions) are necessary. In this analysis the six load cases
were selected to be axial compression (F
8
), torsion (M
8
),
4-point-bending around an axis parallel (M
7
) and an axis
perpendicular (M
6
) to the xateur plane, cantilever ben-
ding around an axis parallel (F
6
, M
7
) and an axis perpen-
dicular (F
7
, M
6
) to the xateur plane (Fig. 3). The loads
and movements were oriented according to a coordinate
system displayed in Fig. 1.

F
V
F
W
F
X
M
V
M
W
M
X

G
"

*
V
*
W
*
X

G
i"1,
2
, 6. (1)
The units of the stiness values are Nmm\ for s

to
s

, s

to s

, s

to s

, Ndeg\ for s

to s

, s

to s

,
s

to s

, Nmmm\ for s

to s

, s

to s

, s

to
s

and Nmdeg\ for s

to s

, s

to s

, s

to s

.
All experiments except those in torsion were per-
formed on a materials testing machine (Zwick 1454,
Zwick GmbH, Ulm, Germany) with specially designed
jigs to allow compression, 4-point-bending and cantilever
bending. Torsional testing was performed on a custom
built device. The inter-fragmentary movements were re-
corded during loading and unloading at 2 mmmin\ or
2 min\. Maximum loads were chosen to be within the
linear behavior of the external xateur constructs (Table
1). Inter-fragmentary movements were recorded with
a personal computer and custom software (Wilke et al.,
1994).
To analyze the inuence of the free length of the
Schanz screws, the carbon rods were moved incremen-
tally 2.5 mm towards and away from the rod axis cre-
ating distances between the xateur body and rod axis of
59.0, 56.5, 54.0, 51.5 and 49.0 mm (Fig. 1).
To predict inter-fragmentary movement the bony load
state was derived from a 3-D model of the hind limb of
a sheep (Duda and Claes, 1996). It was assumed as
a simplication that the bony loads acting on an intact
tibia are similar to those in one that has been osteo-
tomized and stabilized with an external xateur. To
be able to compute inter-fragmentary movements, the
248 G.N. Duda et al. / Journal of Biomechanics 31 (1998) 247252
Fig. 1. ASIF xateur externe construct as used in in vivo measurements
of inter-fragmentary movements in diaphyseal fractures of sheep tibiae
(Stu rmer, 1988). All dimensions are in mm. The origin of the coordinate
system used is at the center of the fracture gap. The x-axis is pointing
to ventral, and y-axis to lateral (left hind limb) and the z-axis is pointing
to proximal. The free length of the Schanz screws is increased from54 to
56.5 and 59 mm and decreased to 51.5 and 49 mm total length each.
inverse of the stiness matrix [Eq. (1)] had to be cal-
culated. Inter-fragmentary movements were calculated
from the inverse stiness matrix multiplied by the dia-
physeal load state of the sheep tibia (Duda and Claes,
1996; F
V
: 0 N shear to ventral, F
W
: !50 Nshear to lateral,
F
X
: !1377 N axial compression, M
V
: !8.8 Nmbending
to lateral, M
W
: 2.5 Nm bending to ventral, M
X
: 0 Nm
torsional moment).
Finally, the inuence of the xateur mounting plane
was analyzed. By mathematically rotating the inverse
stiness matrix around the long axis of the tibia (z-axis),
dierent xateur mounting planes could be simulated.
Starting with a ventral mounting the xateur was rotated
in steps of 10 to a lateral position.
3. Results
With the free length of 54 mm the ASIF external x-
ateur had a stiness of 425.5 Nmm\ in axial compres-
sion, 1.3 Nmdeg\ in torsion, 7.7 Nmdeg\ in bending
perpendicular to the xateur plane and 36.4 Nmdeg\
in bending parallel to the xateur plane. The complete
stiness matrix for a ventrally mounted xateur was:
S"

2298.9 41.6 !36.5 11.5 443.4 !49.4

!70.2 743.8 10.2 !50.6 !62.9 2.1
!735.8 !64.9 425.5 90.6 !225.7 10.5
0.3 2.9 0.5 7.7 !2.2 !0.2
21.7 4.5 !3.8 3.5 36.4 !3.0
!2.2 1.6 0.3 !0.6 !1.7 1.3

.
(2)
The stiness values changed if the free length of the
Schanz screws was increased or decreased, primarily in
the diagonal values of the matrix. To simplify the data,
only the diagonal values of the stiness matrix were given
as a function of the distance between xateur body and
bone axis (Fig. 4). Besides the non-linear decrease in axial
stiness with increasing distance the shear and bending
stinesses perpendicular to the xateur plane decreased.
Little eect could be found for the shear and bending
stinesses parallel to the xateur plane and the torsional
stiness.
Using the most rigid xateur construct (free length
49 mm) and the load state in the tibia diaphysis the
inter-fragmentary movement was 0.46 mm in compres-
sion and 0.22 mm laterally (Fig. 5, ventral mounting). If
the xateur was not completely aligned with the bony
axis but an additional parallel shift laterally of 1 mm was
introduced, the bending moment around the ventral
oriented axis and the torsional moments slightly in-
creased. This resulted in an inter-fragmentary movement
of 0.45 mm in compression and 0.19 mm laterally.
The inuence of the orientation of the xateur plane is
given in Fig. 5 with inter-fragmentary movements for all
positions between a ventral and lateral mounting of the
external xateur. The inter-fragmentary movement in
compression was reduced to nearly zero from a ventral to
a lateral mounting. In contrast, shear movements showed
their maximum for a pure lateral mounting of the
external xateur.
4. Discussion and conclusions
A method was developed to determine the 3-Dstiness
of fracture xation devices. Based on the 3-D stiness,
inter-fragmentary movement was predicted as a function
of fracture location and xateur mounting.
In contrast to single stiness values, the 3-D stiness
description allows dierentiation between more rigid and
less rigid load planes. From the diagonal stiness values
the selected xateur construct appears rather rigid in
axial compression but less rigid under shear and bending
perpendicular to the xateur plane. If the distance
between xateur body and bone axis is increased, the
G.N. Duda et al. / Journal of Biomechanics 31 (1998) 247252 249
Fig. 2. Test setup to measure the axial stiness with the goniometer system attached to the distal and proximal Schanz screws of the xateur. The 3-D
oset of the measurement points of the goniometer system were used to transform the data into relative displacements at the centre of the fracture gap
(Fig. 1).
construct stiness rapidly decreases both in the axial and
lateral (shear) directions (Fig. 4).
In vivo measurements and analytical studies report
that bones are mainly axially loaded with only moderate
bending (Duda et al., 1997; Schneider et al., 1990). The
load magnitudes selected to predict inter-fragmentary
movement (Duda and Claes, 1996) are comparable in
magnitude with those reported from a 2-D analysis (Hut-
zschenreuter et al., 1993). Hutzschenreuter et al. report
tibial compression of 1090 N (the present study 1377 N)
and bending around a laterally oriented axis of 9.4 Nm
(here 8.8 Nm).
Although the loading is mainly axial, the computed
inter-fragmentary movements are rather complex
(Fig. 5). Nonetheless, the inter-fragmentary movements
recorded in an in vivo study on tibial fractures in sheep
(Stu rmer, 1988) are comparable to those calculated in
this study. Stu rmer reports for a diaphyseal fracture
during stance phase axial movement of 0.46 mm (here
0.46 mm) and lateral movement of 0.19 mm (here
0.22 mm).
From these similarities it appeared warranted to ex-
tend the presented method to predict inter-fragmentary
movements under various xateur congurations. As an
illustration, the inuence of xateur mounting plane on
inter-fragmentary movement was selected. A simple
method to achieve optimal mechanical conditions at the
fracture site (0.21.0 mm axial and minimal shear
movements) was the modication of the xateur
mounting plane. In the present xateur conguration,
250 G.N. Duda et al. / Journal of Biomechanics 31 (1998) 247252
Fig. 3. The six linear independent load cases used to determine the stiness matrix. From left to right: Axial compression, torsion, 4-point-bending
parallel and perpendicular to the xateur mounting plane, cantilever bending parallel and perpendicular to the xateur mounting plane.
Table 1
Loads used to measure the inter-fragmentary movement for the six
independent load cases. Coordinates are according to Fig. 1: X to
ventral, Y to lateral, Z to proximal. Origin is the center of the fracture
gap
F
V
F
W
F
X
M
V
M
W
M
X
(N) (N) (N) (Nm) (Nm) (Nm)
Axial compression 0 0 !500 0 0 0
Torsion 0 0 0 0 0 2.5
4-point-bending,
parallel 0 0 !200 0 !10.1 0
4-point-bending,
perpendicular 0 0 !200 !10.1 0 0
Cantilever bending,
parallel 200 0 0 0 19.6 0
Cantilever bending, 0 !90 0 8.8 0 0
perpendicular
axial movement was well within the range considered
appropriate for optimal fracture healing (Claes et al.,
1995) but the shear movement was never adequately
minimized in any mounting plane (Fig. 5). A laterally
mounted xateur with minimal axial movement and
maximum shear would probably provide the least eec-
tive mechanical environment for fracture healing. The
data support rather a 20 ventro-lateral mounting, if
allowed by anatomical constraints.
The fracture model used in this study assumes no load
transfer across the fracture gap and, therefore, represents
a worst case scenario. The model would have to be
adapted to other fracture modes by adjusting the xation
stiness appropriately. Such adaptations as well as pre-
dictions of inter-fragmentary movements are only man-
ageable by 3-D descriptions of the xation stinesses.
Fig. 4. The inuence of the free length of the screws on the diagonal
values of the stiness matrix. #F
V
/u
V
: ventral shear stiness, #F
W
/u
W
:
lateral shear stiness, !F
X
/u
X
: compression stiness, #M
V
/*
V
: ventral
ending stiness, #M
W
/*
W
: lateral bending stiness, #M
X
/*
X
: torsional
stiness.
The complete description provides a tool to investi-
gate inter-fragmentary movement under any possible
load condition in vivo. With knowledge of the optimal
axial and shear movements it further allows the appro-
priate selection of xateur congurations to control
inter-fragmentary movement.
G.N. Duda et al. / Journal of Biomechanics 31 (1998) 247252 251
Fig. 5. Inter-fragmentary movement of a mid-shaft tibia osteotomy
stabilized with an ASIF xateur mounted in a ventral, ventro-lateral or
lateral position. #u
V
: shear to ventral, #u
W
: shear to lateral, !u
X
:
compression.
Acknowledgements
The authors would like to thank Patricia Horny, De-
partment of Unfallchirurgische Forschung und Bio-
mechanik, University of Ulm, for helping with the graphs
and gures and Dr K. Wenger, Department of Unfal-
lchirurgische Forschung und Biomechanik, University of
Ulm, for editing.
References
Claes, L., Wilke, H.-J., Augat, P., Ru benacker, S., Margevicius, K.J.,
1995. Eect of dynamization on gap healing of diaphyseal fractures
under external xation. Clinical Biomechanics 10, 227234.
Cunningham, J.L., Evans, M., Kenwright, J., 1989. Measurement of
fracture movement in patients treated with unilateral external skel-
etal xation. Journal of Biomedical Engineering 11, 118122.
Duda, G.N., Chao, E.Y.S., Schneider, E., 1997. Internal forces and
moments in the femur during walking. Journal of Biomechanics 30,
933941.
Duda G.N., Claes, L., 1996. Prediction and control of 3-D interfrag-
mentary movement in fracture healing. In Proceedings of the Inter-
national Society of Fracture Repair. Ottawa, p. 24.
Gardner, T.N., Hardy, J.R.W., Evans, M., Richardson, J.B., Kenwright,
J., 1996. The static and dynamic behaviour of tibial fractures due to
the procedure of dynamisation. Clinical Biomechanics 11, 425430.
Goodship, A.E., Kelly, D.J., Rigby, H.S., Watkins, P.E., Kenwright, J.,
1988. The eect of dierent regimes of axial micromovement on the
healing of experimental tibial fractures. In: Bergmann G., Ko lbel R.,
Rohlmann A. (Eds.), Biomechanics: Basic and applied research, Mar-
tinus Nijho Publishers, Dordrecht, pp. 441446.
Goodship, A.E., Kenwright, J., 1985. The inuence of induced micro-
movement upon the healing of experimental tibial fractures. Journal
of Bone and Joint Surgery 67, 650655.
Goodship, A.E., Watkins, P.E., Rigby, H.S., Kenwright, J., 1993. The
role of xator frame stiness in the control of fracture healing. An
experimental study. Journal of Biomechanics 26, 10271035.
Homann, R., McKellop, H., Sarmiento, A., Lu, B., Ebramzadeh, E.,
1991. Three-dimensional measurement of fracture gap motion. Bio-
mechanical study of experimental tibial fractures with anterior clasp
xator and ring xator. Unfallchirurg 94, 395400.
Hutzschenreuter, P.O., Sekler, E., Faust, G., 1993. Loads on muscles,
tendons and bones in the hind extremities of sheep a theoretical
study. Anat. Histol. Embryol. 22, 6782.
Kenwright, J., Goodship, A.E., 1989. Controlled mechanical stimula-
tion in the treatment of tibial fractures. Clinical Orthopedics 241,
3647.
Kenwright, J., Richardson, J.B., Cunningham, J.L., White, S.H., Good-
ship, A.E., Adams, M.A., Magnussen, P.A., Newman, J.H., 1991.
Axial movement and tibial fractures. A controlled randomised trial of
treatment. Journal of Bone and Joint Surgery 73, 654659.
Kristiansen, T., Fleming, B., Neale, G., Reinecke, S., Pope, M.H., 1987.
Comparative study of fracture gap motion in external xation. Clini-
cal Biomechanics 2, 191195.
Lippert, F.G., Hirsch, C., 1974. The three-dimensional measurement of
tibia fracture motion by photogrammetry. Clinical Orthopedics 105,
130143.
Molster, A.O., Gjerdet, N.R., 1984. Eects of instability on fracture
healing in the rat. Acta Orthopedica Scandinavica 55, 342346.
Molster, A.O., Gjerdet, N.R., Raugstad, T.S., Hvidsten, K., Alho, A.,
Bang, G., 1982. Eect of instability on experimental fracture healing.
Acta Orthopedica Scandinavica 53, 521526.
Schenk, R.K., Mu ller, J., Willenegger, H., 1986. Experimentell-his-
tologischer Beitrag zur Entstehung und Behandlung von Pseudar-
throsen. Hefte zur Unfallheilkunde 94, 1524.
Schneider, E., Michel, M.C., Genge, M., Perren, S.M., 1990. Loads
acting on an intramedullary femoral nail. In: Bergmann G., Graichen
F., Rohlmann A. (Eds.), Implantable Telemetry in Orthopaedics,
Freie Universita t Berlin, Berlin, pp. 221227.
Stu rmer, K.M., 1988. Histologie und Biomechanik der Frakturheilung
unter den Bedingungen des Fixateur externe. Hefte zur Unfall-
heilkunde 200, 233242.
Wilke, H.-J., Ostertag, G., Cleas, L., 1994. Three-dimensional
goniometer linkage system for the analysis of movements with six
degrees of freedom. Biomedizinische Technik (Berlin) 39, 149155.
Yamagishi, M., Yoshimura, Y., 1955. The biomechanics of fracture
healing. Journal of Bone and Joint Surgery 37, 10351068.
252 G.N. Duda et al. / Journal of Biomechanics 31 (1998) 247252