Arch Orthop Trauma Surg (2002) 122 : 88–92
DOI 10.1007/s004020100324
C L I N I C A L A N D E X P E R I M E N TA L F O R U M
Jin-Won Yang · Kyung-Hoi Koo · Min-Cheol Lee ·
Philip Yang · Myunggyu D. Noh · Shin-Yoon Kim ·
Kang-Il Kim · Yong-Chan Ha · Man-Soo Joun
Mechanics of femoral head osteonecrosis
using three-dimensional finite element method
Received: 15 December 2000 / Published online: 5 September 2001
© Springer-Verlag 2001
Abstract A three-dimensional finite element model of a
femoral head was developed using a surface modeling
technique. The distribution of the stress index S (S = ef-
fective stress / yield strength, σ–/σY) in various sizes of seg-
mental osteonecrosis was assessed. The stress index of the
femoral head was within physiological limits when the
necrotic angle was less than 110°. Within both the sub-
chondral region and the deep necrotic region adjacent to
the necrotic-viable interface, values of the stress index sig-
nificantly higher than the normal physiological level (>0.1)
appeared when the necrotic angle was 110° or more.
In the analysis of 28 osteonecrotic femoral head speci-
mens, fracture appeared in two major locations: the deep
necrotic region near the underlying necrotic-viable interface
(19 femoral heads) and the subchondral region (7 femoral
heads). In 2 femoral heads, the fracture involved both re-
gions. Both sites of fracture coincided with the region of
stress index greater than 0.1 in the finite element model
study. These results may provide baseline information for
predicting the collapse of the femoral head and determin-
ing the treatment modality of early stage osteonecrosis.
J.-W. Yang · K.-H. Koo (✉) · P. Yang · K.-I. Kim · Y.-C. Ha
Department of Orthopaedic Surgery,
Gyeong-Sang National University School of Medicine,
90 Chilam-dong, Chinju 660–702, South Korea
e-mail: khkoo@nongae.gsnu.ac.kr,
Tel.: +82-55-7508107, Fax: +82-55-7534789
M.-C. Lee · M.-S. Joun
Regional Research Center for Aircraft Parts Technology,
Gyeong-Sang National University, 900 Gajwa-dong,
Chinju 660–701, South Korea
M.D. Noh
Department of Mechatronics Engineering,
Chungnam National University, 6 Moonhwa-dong,
Taejon 301–131, South Korea
S.-Y. Kim
Department of Orthopaedic Surgery, Kyung-Pook National Uni-
versity School of Medicine, 101 Dongin-2-ga,
Taegu 700–422, South Korea
Keywords Femoral head · Osteonecrosis · Finite
element method · Extent
Introduction
Previous clinical studies have indicated that the extent and
site of necrosis affect the development of the collapse of
the femoral head in osteonecrosis [2, 10, 11, 12, 13, 14].
A small area of osteonecrosis does not lead to collapse,
whereas a medium to large area of osteonecrosis frequently
does. A nonlinear finite elementary model (FEM) study
indicated that the stress pattern changes in osteonecrotic
femoral heads are particularly sensitive to the size and
shape of the necrotic lesion and to its location relative to the
weight-bearing portion [6]. However, previous three-di-
mensional FEM studies did not consider changes in stress
and strength distribution according to the size of the
necrotic portion, and the border of the necrotic portion
was not precisely reflected in finite element mesh gen-
eration [3, 5, 7]. In addition, a longitudinal compres-
sive load was applied to a point on the femoral head.
This assumption will likely lead to erroneous results be-
cause, in reality, the compressive load is distributed on an
area of the weight-bearing portion of the femoral head.
The present study investigated the stress and strength
distribution on various sizes of necrotic areas. The contact
region between the femoral head and the acetabulum was
treated as a contact problem, and the necrotic boundary was
clearly expressed in the analysis model. We quantified the
stress index, volume ratio of the dangerous portion, and
mean stress index of volume. The results of the analysis
were compared with the fracture lines of osteonecrotic
femoral heads. Also, the relationship between the extent
of the necrosis and the collapse of the femoral head was
evaluated.
The results of the present study may be useful to predict
the fate of the femoral head and to determine the treatment
modality for early stage osteonecrosis. This study also sup-
plies the baseline data to validate currently used joint-pre-
serving procedures, such as core decompression and vari-
ous osteotomy methods.

Materials and methods
A three-dimensional FEM was constructed using a surface model-
ing technique. Because this study focuses on the prediction of col-
lapse due to necrosis, finite element mesh generation of the necrotic
portion was very detailed, whereas geometric approximation was
used for the other parts. The femoral head was considered a hemi-
sphere with a radius of 25 mm. The central axis of the hemisphere
was assumed to be 45° away from the direction of gravitational
force. The finite element mesh used in the present study was com-
posed of 4367 elements and 4466 nodes (Fig. 1).
The shape and location of necrosis were determined by analyz-
ing magnetic resonance imaging (MRI) scans of 55 patients with
osteonecrosis. We assumed the necrotic portion of the femoral
head was shaped like a cone, and the axis of the cone was the same
as the axis of the femoral head hemisphere. The size of the necro-
sis was defined as having an angle of α and a height of h. We
Fig. 1 Three-dimensional finite element mesh was composed of a
4367 elements and 4466 nodes
Fig. 2 Geometrical definitions of the hip joint include the angle of
necrosis (α);angle between the vertical axis and the central axis of
necrosis (β);angle of the contact region of the acetabulum (γ);angle
between the vertical axis and the axis of loading (δ);height of the
necrotic portion (h)
89
found an approximation formula between the angle of necrosis and
the height of the cone. The height was determined as follows:
α
h = r × sin .
2
The acetabulum was regarded as a rigid body. The interface be-
tween the femoral head and the acetabulum was treated as a con-
tact problem. The angle of the contacting region was assumed to
be 130°. The published data [1, 3, 4, 5, 6, 7, 8, 9] were used for the
mechanical properties of normal and osteonecrotic cancellous bone,
such as stiffness, strength, and Poisson’s ratio, as well as the load-
ing condition. We assumed the entire femoral head to be composed
of an isotropic material and ignored the effects of cortical bone.
Values of 445 MPa and 218 MPa were used for the modulus of
elasticity of the viable portion and necrotic portion, respectively;
0.3 was used for both portions for Poisson’s ratio; and 19.4 MPa
and 5.5 MPa for the yield strength of the viable portion and
necrotic portion, respectively. The cartilage was assumed to sim-
ply transmit the joint loading, passively, to underlying subchondral
bone. A load of 3,000 N was applied 15° apart from the direction of
gravitational force towards the center of the femoral head. To treat
the boundary between the femoral head and acetabulum as a con-
tact problem, the load F was not applied to nodes of the contact re-
gion but to a node at the bottom of the femoral neck. The degree of
freedom in the normal direction was eliminated, and through itera-
tive calculation, the boundary condition was corrected so that tensile
forces were not applicable to contact nodes. Consequently, the load
that is applied to the nodes of the contact region becomes an un-
known variable, and the sum of external forces becomes 3,000 N
directed towards the center of the femoral head (Fig. 2).
We defined the ratio of effective stress (σ– ) and yield strength
(σY) as stress index S (S = effective stress / yield strength, σ–/σY).
Theoretically, yield or fracture starts to occur when the stress in-
dex is larger than 1. However, a fatigue fracture can occur even
when the stress index is less than 1.
The purpose of the present study was to evaluate the risk of
fracture according to the size of the necrotic portion. The volume
ratio of the risky portion, R, was defined as the ratio of the volume
in which the stress index was greater than the maximum stress
index in the normal femoral head, S’ in the hemisphere to the to-
tal volume of the hemisphere (V). This was calculated as follows:
R= V
∫ < S − S ′ > 0 dV
.
V
The singularity function becomes one if s-s’ is positive and be-
comes zero if s-s’ is negative.
The mean stress index of volume (I) was defined as follows:
I= V
∫ SdV .
V
Fig. 3 Stress index distribution of the normal femoral head
90
Fig. 4A–F Stress index distri-
bution according to the necrotic
angle of 70° (A), 90° (B),
110° (C), 120° (D), 140° (E),
and 150° (F)

A total of 12 simulations were performed according to the

necrotic angle (the angle of the necrotic portion in mid-coronal and
mid-sagittal sections) from 60° to 170° at 10° intervals.
Twenty-eight osteonecrotic femoral heads were obtained from
patients undergoing total hip arthroplasty. The osteonecrotic
femoral heads were cut in the mid-coronal plane. Locations of
fracture lines were analyzed and compared with the results of the
FEM study.
Results
The stress index of the normal femoral head ranged from
0.05 to 0.10 and showed a relatively uniform distribution
(Fig. 3). Thus, the maximum stress index value of the nor-
mal femoral head was 0.10. In the osteonecrotic femoral
 91

head, the stress index varied substantially according to the

changes in the area of necrosis.
When the necrotic angle was less than 110°, the stress
index in the necrotic portion was not more than 0.10, which
was the highest physiologic level. When the necrotic an-
gle was 110°, a high stress index of 0.15 appeared in the
deep necrotic portion above the necrotic-viable interface.
When the necrotic angle was 120°, the maximum stress
index reached 0.2, and a high stress index of 0.15 appeared
in the subchondral area. When the necrotic angle was 150°,
the maximum stress index reached 0.3 (Fig. 4). Hence the
stress index was not more than 1.0 even when the necrotic
area was quite large. This finding reflects that the collapse
of the necrotic femoral head is caused by a fatigue frac-
ture.
Because the maximum stress index value of the normal
femoral head was 0.10, the portion in which the stress in-
dex (S’) was 0.15 or more was considered as the danger-
ous portion in which a fatigue fracture might occur. When
there was no necrosis, the volume ratio of the dangerous
portion was 0.0. As the necrotic portion increased, the vol-
ume ratio increased almost linearly and reached 0.1. How-
ever, the mean stress index of volume exhibited two dif-
ferent patterns. When the necrotic angle was less than
110°, the mean stress index of volume increased slowly.
When the necrotic angle was more than 110°, the mean
stress index of volume increased abruptly. Thus, the vol-
Fig. 6A,B Fracture sites of necrotic femoral heads are shown at
ume ratio of the dangerous portion showed an abrupt in- the subchondral region (A) and at the deep necrotic portion adja-
crease as the necrotic angle widened (Fig. 5). This result cent to the necrotic-viable junction (B)
was related to the finding that changes in the stress index
distribution pattern occurred at approximately 110°.
The results of analysis of 28 osteonecrotic femoral head interface (19 femoral heads) and in the subchondral area
specimens demonstrated two major locations of fracture: (7 femoral heads). In 2 femoral heads, the fracture occurred
in the deep necrotic portion adjacent to the necrotic/viable in both regions (Fig. 6). These data were consistent with the
results of the FEM study. Those two sites of fracture co-
incided with the region of stress index of more than 0.1 in
the FEM study, which appeared in the subchondral and
the deep necrotic regions when the angle of the necrotic
portion was larger than 110°.

Discussion

This study indicates that the regions at risk for fracture

Fig. 5 Volume ratio of dangerous portion (R) and mean stress in-
dex of volume (I)
with a stress index more than 0.1 appear in the subchon-
dral area and deep necrotic region near the necrotic-viable
interface when the necrotic angle is larger than 110°. In
segmental osteonecrosis with a necrotic angle less than
110°, the stress index in the necrotic portion was not more
than the physiological level (0.10 or less), and the risk is
low that the femoral head will fracture and subsequently
collapse. In necrosis with a necrotic angle of 110° or larger,
the stress index was above the physiological level (0.15 or
more) at the deep necrotic portion and subchondral re-
gion, and the femoral head is at risk for fracture and col-
lapse.
Previous clinical studies have indicated that the extent
and site of the necrotic portion affect the development of
collapse of the femoral head in osteonecrosis [2, 10, 11,
92
12, 13, 14]. A small area of osteonecrosis does not lead to
collapse, whereas a medium to large area of osteonecrosis
frequently does. A two-dimensional FEM study showed
that the size of the necrotic portion affects the development
of collapse of an osteonecrotic femoral head [6]. However,
previous three-dimensional FEM studies did not consider
changes in stress and strength distribution according to the
size of the necrotic portion, and the border of the necrotic
portion was not precisely considered in finite element mesh
generation [3, 5, 7].
The present study approached stress analysis of the
necrotic femoral head quantitatively with the recognition
of the problems presented in previous studies. The os-
teonecrotic femoral head was evaluated using a three-di-
mensional finite element analysis to overcome the limita-
tion of the two-dimensional finite element method. The
contact region between the femoral head and acetabulum
was treated as the contact problem. The load was not ap-
plied to nodes of the contact region but to a node on the
opposite side, and the necrotic boundary was clearly ex-
pressed in the analysis model. For quantification of the re-
sults, we quantified the stress index, volume ratio of the
dangerous portion, and mean stress index of volume.
The result of the present finite element study involving
various areas of necrosis may be useful to predict the fate
of the femoral head and to determine the treatment modal-
ity in early stage osteonecrosis. Any surgical intervention
to prevent collapse should be limited to those hips at risk
of fracture, because a femoral head with a small area of
osteonecrosis is at low risk for collapse. This study also
supplies the baseline data to confirm the validity of cur-
rently used joint-preserving procedures, such as core de-
compression and various osteotomy methods.
References
1. Baker KJ, Brown TD, Brand RA (1989) A finite-element analy-
sis of the effects of intertrochanteric osteotomy on stress in
femoral head osteonecrosis. Clin Orthop 249: 183–198
2. Bonfiglio M, Bardenstein MB (1958) Treatment by bone-graft-
ing of aseptic necrosis of the femoral head and non-union of
the femoral neck (Phemister technique). J Bone Joint Surg Am
40: 1329–1346
3. Brown TD, Hild GL (1983) Pre-collapse stress redistributions
in femoral head osteonecrosis – a three-dimensional finite ele-
ment analysis. J Biomech Eng 105: 171–176
4. Brown TD, Way ME, Fergusson AB Jr (1980) Stress transmis-
sion anomalies in femoral heads altered by aseptic necrosis.
J Biomech 13: 687–699
5. Brown TD, Way ME, Fergusson AB Jr (1981) Mechanical char-
acteristics of bone in femoral capital aseptic necrosis. Clin Or-
thop 156: 240–247
6. Brown TD, Mutschler TA, Ferguson AB Jr (1982) A non-lin-
ear finite element analysis of some early collapse processes in
femoral head osteonecrosis. J Biomech 15: 705–715
7. Brown TD, Baker KJ, Brand RA (1992) Structural conse-
quences of subchondral bone involvement in segmental osteo-
necrosis of the femoral head. J Orthop Res 10: 79–87
8. Brown TD, Pedersen DR, Baker KJ, Brand RA (1993) Me-
chanical consequences of core drilling and bone grafting on os-
teonecrosis of the femoral head. J Bone Joint Surg Am 75: 1358–
1367
9. Huiskes R, Strens, PH, Van Heck J, Slooff TJ (1985) Interface
stresses in the resurfaced hip. Finite element analysis of load
transmission in the femoral head. Acta Orthop Scand 56: 474–478
10. Kerboul M, Thomine J, Postel M, Merle d’Aubigne R (1974)
The conservative surgical treatment of idiopathic aseptic necro-
sis of the femoral head. J Bone Joint Surg Br 56: 291–296
11. Koo KH, Kim R (1995) Quantifying the extent of osteonecro-
sis of the femoral head. A new method using MRI. J Bone Joint
Surg Br 77: 875–880
12. Koo KH, Kim R, Ko GH, Song HR, Jeong ST (1995) Preventin
g collapse in early-stage osteonecrosis of the femoral head: a
randomized clinical trial of core decompression. J Bone Joint
Surg Br 77: 870–874
13. Shimizu K, Moriya H, Akita T, Sakamoto M, Suguro T (1994)
Prediction of collapse with magnetic resonance imaging of
avascular necrosis of the femoral head. J Bone Joint Surg Am
76: 215–223
14. Sugano N, Takaoka K, Ohzono K, Matsui M, Masuhara K, Ono
K (1994) Prognostication of nontraumatic avascular necrosis of
the femoral head. Clin Orthop 303: 155–164