Correspondence to: JHB.

Current Orthopaedics (2001) 15, 57d67
^

2001 Harcourt Publishers Ltd

doi:10.1054/cuor.2001.0140, available online at http://www.idealibrary.com on

LABORATORY RESEARCH

Experimental avascular osteonecrosis

J. H. Boss and I. Misselevich

MD Department of Pathology, Bnaizion Medical Center, P.O. Box 4940, Haifa 31048, Israel

INTRODUCTION

Designing suitable, fitting, and readily reproducible
experimental models of osteonecrosis is more than an
academic exercise. Mont and Hungerford argue that
treatment options of patients with osteonecrosis are
difficult to explore because novel therapeutic modalities
cannot be evaluated in animals with an appropriately
congenial, experimentally induced counterpart of the
human disease.

1

A suitable animal model ought to

duplicate

aetiologically

and

pathogenetically

the

‘circulatory deprivation’ settings implied by clinicians’
time-honoured practice to synonymously use the
epithets of ‘avascular’ osteonecrosis and ‘idiopathic’
osteonecrosis.

Interruption of the blood supply to a tissue for

a specific period of time results in cell death. That
ischaemic events precipitate necrosis of the fibrous,
haematopoietic, fatty, and bony tissues of the skeleton is
inherent in the ‘bone infarction’ paradigm. Irrespective of
the site at which the blood flow is initially disrupted, i.e.
the sinusoidal, arterial, capillary or venous network, the
circulation in the arteries is utlimately compromised. To
duplicate avascular osteonecrosis of an adult patient’s
femoral head, severing the blood supply or blood
drainage, ought to result in necrosis of an animal’s
femoral head. To emulate a child’s Perthes disease,
disrupting the blood supply or blood drainage ought to
produce necrosis of an animal’s femoral epiphysis in the
presence of physeal cartilage.

2

Currently studied models

properly mimic one aspect or another of the disorder.
Alas, no model satisfactorily mirrors all the patho-
physiologic and microscopic traits of osteonecrosis as
encountered in patients. Equating the histological
features of necrotic bone with those of say, a renal or
a splenic infarct, authors in the 1960s spoke of ‘bone
infarction’ where contemporary orthopaedists would
apply the label of osteonecrosis.

3

The healing processes governing the substitution of

the dead tissues by the viable tissues in an avascular
necrotic bone and a fractured bone share common

features, beginning with the inflammatory response,
continuing with angiogenesis and re-establishment of
the blood flow, fibrogenesis, osteogenesis and variable
chondrogenesis, and winding up with modelling of
the newly synthesized bone with restoration of its
mechanical properties, albeit incomplete in the ordinary
case of osteonecrosis. In contrast to a healed bone after
a conventional fracture, the original architecture and
strength of healed bones after an infarction are typically
not restored, such that deformities remain as a rule.
Skeletal segments with a poorly developed collateral
circulation are at an increased risk of ischaemic injury.
Hence, the parts of the skeleton adjacent to joints
are primarily involved, carrying with it a probability of
an evolving osteoarthritis. This diaphyseal cortex
infrequently endures ischaemic necrosis because of its
redundant blood supply.

2,4

Operationally defined, osteonecrosis is characterized

by ‘optically empty’ lacunae (Fig. 1). Ischaemic and non-
ischaemic osteonecrosis have to be differentiated from
each other. By definition, the cause of idiopathic
osteonecrosis is unclear. Symptomatic osteonecrosis is
caused by any of a number of catastrophic occurrences:
(a) disrupted arterial input secondary to luminal
obstruction by internal obliteration or external com-
pression; (b) hindered circulation within the intraoss-
eous arterioles; (c) arterial insufficiency resulting from
a deranged venous drainage; (d) primary extravascular
disorders with a decreased venous outflow, secondarily
leading to decreased arterial blood flow; (e) Deep vein
thrombosis; (f ) various cytotoxic agents. Co-operating
with one another, and further miscellaneous factors,
these conditions are not mutually exclusive. Indeed,
‘avascular’ osteonecrosis comprises a multifactorial,
heterogeneous group of disorders sharing the final
common pathway of circulatory failure.

1,2,5,6

Except for so-called ‘cytotoxic’ bone necrosis (causally

related to the effects of chemotherapy, radiotherapy and
physical or chemical injuries), osteonecrosis is the result
of a disparity between the osteocytes’ need and supply
of oxygen. In this context, arterial blockage is more
common than obstruction of the venous drainage, the
total volume of the osseous veins being much greater

Figure 1

Osteonecrosis on the 5th day after the disruption of

all the blood vessels supplying and draining a rat’s femoral head.
The lacunae of the vast majority of the osteocytes are optically
empty (arrow heads). The intertrabecular spaces are partly occu-
pied by amorphous or granular, eosinophilic matter mixed with
haematoxylinophilic, karyopyknotic and karyorrhectic particles
(arrows). Elsewhere, fibrous tissue (FT) has replaced the necrotic
tissue in the intertrabecular spaces. (Reproduced from Experi-
mental and Molecular Pathology with permission of Academic
Press).

Figure 2

A well-vascularized, loosely textured fibrous tissue,

which has replaced the necrotic fatty-haematopoietic marrow,
occupies the intertrabecular spaces. Note the optimally empty
osteocytic lacunae (arrow heads) of the osseous trabeculae of the
necrotic femoral head. The intertrabecular spaces are occupied
by young fibrous tissue. (Reproduced from Experimental and
Molecular Pathology with permission of Academic Press).

than that of the arteries.

7d9

Winet et al. argue that the

described experimental models do not satisfy the
requirements of a faithful reproduction because they
fail to duplicate one aspect or another of the human
counterpart. Man-made artifacts in every intervention
are all too obvious. In addition, the healing potential of
animals’ wounded tissues, and of small laboratory animals
in particular, surpass many times the reparative capacity
of human tissues, playing down to some degree the
merit of the researched models.

2,10

A crucial issue is the

duration of ischaemia critical to irreversibly damage the
cells. Osteocytes may be less sensitive to an ischaemic
insult than other cell types. An adult dog’s osteocytes
survive, at least partly, many hours of ischaemia as
attested by the results of reimplantation of a limb after
3 days of ex vivo maintenance at 43C. True, bone cells of
young animals are less resistant to oxygen deprivation
than the osteocytes of adult animals. Nevertheless, bone
growth also goes on in puppies’ limbs replanted after
having been preserved ex vivo for 3 days at 43C.

11,12

SURGICALLY INDUCED
OSTEONECROSIS

A sudden disruption of the arterial blood supply leads to
necrosis of osseous tissues. Large diaphyseal cortical
sections

undergo

necrosis

after

reaming-induced

devascularization of the medullary cavity of adult dogs’
femurs. That all dead bone is resorbed and replaced
by newly synthesized bone within 6 months of the
operation attests to the extraordinary regenerative

potential of animals’ tissues. Quantification of the
turnover of rats’ cortical bones prelabelled with

3

H-

tetracycline,

3

H-proline,

and

45

Ca

indicates

that

disruption of the blood supply without violating the
osseous integrity leads to enhanced bone resorption and
formation during the first two post-operative months.
The amounts of band collagen and calcium do not change
significantly during this time period. Therefore, the
increased osteolysis is entirely compensated for by an
equivalent increased osteogenesis.

13,14

Necrosis

follows

the

removal,

destruction

or

obstruction of the retinacular vessels of adult dogs’
femoral head. Capillary-sized anastomoses connect the
epiphyseal and metaphyseal circulation with each other.
Stripping the retinacular vessels devascularizes the bone.
Reaming of the femoral head does not by itself exert
such an effect. A maximal rate of blood flow reduction is
attained by reaming combined with retinacular vessel
stripping. The collateral epiphyseal-metaphyseal anasto-
moses in dogs with chronic arthritis make their femoral
head less vulnerable to disruption of the retinacular
vessels.

15

Granulation tissue purportedly replaces the haema-

topoietic and fatty marrow in ‘fresh’ osteonecrosis of the
femoral head. This premise is inaccurate, to say the least,
existence of granulation tissue excluding a recently
passing necrotic event. It takes a few days for inflam-
matory cells to infiltrate throughout the debris and for
the fibroblasts and vessels to circumferentially penetrate
into and replace the necrotic cells and connective tissue
fibres occupying the intertrabecular spaces of a dead
bone (Fig. 2). Mesenchymal progenitor cells, accom-
panying, as well as migrating from, the invading vessels,
differentiate

into

preosteoblasts.

Having

matured

58

CURRENT ORTHOPAEDICS

Figure 3

Undifferentiated mesenchymal cells (arrows) accom-

pany the vessel (V) growing into an intertrabecular space. The
cuboidal osteoblasts (arrow heads) are aligned along the osteoid
matrix, deposited on the necrotic bony trabeculae. (Reproduced
from Experimental and Molecular Pathology with permission of
Academic Press).

into the osteoblasts, the cells synthesize osteoid matrix,
which is deposited appositionally aside the necrotic
bone (Fig. 3) and intramembraneously amid the newly
formed fibrous tissue, the descendant of the granulation
tissue. The extent of revascularization and the intensity
of the reparative processes correlate with each other.
Microangiographically, no vessels are discernible during
the early stage of osteonecrosis, but they become
apparent with emerging anastomoses between the epi-
physeal and the metaphyseal circulations. As regards
timing, all necrotic soft and hard tissues are resorbed and
replaced by newly formed fibrous, fatty, haematopoietic
and osseous tissues within 6 weeks or so of vascular
deprivation-induced necrosis of the rat’s femoral
head.

10,16

Osteocytes within a 0.1 mm-wide gap of titanium

chambers, implanted into a rabbit’s tibia, die after 20-min
occlusion of the common iliac artery. After occluder
removal followed by a fleeting overshoot of the blood
flow, the circulation reverts to baseline levels by the
7th postoperative week. By way of comparison,
revascularization is realized 4 months after disruption of
the blood supply of dogs’ femoral heads. The ingrowing,
well-vascularized fibrous tissue replaces the necrotic
tissues. The thrombi, which have plugged the venules and
capillaries, are cleansed during the reperfusion phase.
The newly delivered oxygen interacts with the xanthine
oxidase released from the neutrophils. Endothelial cell-
adherent leukocytes and xanthine oxidase-dependent
intermediaries (principally free radicals) injure the
vessels. Having been set free from disrupted lipocytes,
fatty constituents penetrate into blood sinusoids and
exert further noxious effects on all nearby tissues,
including the vessels. These processes, the outcome
of an at least 10-h-lasting ischaemic period, initiate

a secondary ischaemic insult, precipitating the so-called
‘reperfusion injury’. Shortly after blood supply disrup-
tion, capillaries sprout into rabbits’ necrotic femoral
heads. Branching, they form irregular networks of
capillary sized vessels. With time, they enlarge and
transform into arteries, which extend throughout the
femoral head.

1

In rats, revascularization of an avascular

necrotic femoral head is initiated by the end of the
1st postoperative week, resorption of the necrotic bone
is advanced by the 2nd week, appositional and intra-
membranous osteogenesis flourishes by the 3rd week,
vascular perfusion exceeds the normal levels by the 2nd
month, and the circulation normalizes shortly thereafter.
To reiterate, cells restoring skeletal integrity emanate
not only from the granulation tissue but also from the
progenitor cells carried in vessels. The cells migrating
from the vessels mature into macrophages, osteoclasts,
fibroblasts and osteoblasts.

2,17d20

Arterial and venous occlusion by a pneumatic

tourniquet, applied to a rabbit’s mid-thigh, leads to bone
necrosis distal to the vascular obstruction site. Similarly
in rats, reperfusion injury with secondary ischaemia
follows termination of the vascular occlusion. Creeping
substitution of the necrotic debris by fibrous and bony
tissues accompanies renewal of the circulation. The
extent of bone loss is exposure time-dependent, more
osseous tissues being resorbed in rabbits subjected to
a 6-h than a 2-h period of vascular occlusion. By the

99 mm

technetium antimony colloid technique, radioactivity

is not detected on the first day after disrupting the supply
of blood to the femoral head by a subcapital osteotomy
and section of the ligamentum teres. Yet bone imaging
with

99 mm

technetium methylene diphosphonate displays

no deviation from the norm for the first two post-
operative days. This apparent contradiction reflects the
difference between marrow and bone scans and attests
to the earlier loss of function of the haematopoietic cells
than of the osteocytes in the avascular environment.

2,21

The intra-articular pressure depends on the amount of

fluid in the joint cavity and the position of the joint. For
instance, the pressure in a hip joint is highest in the
externally or internally rotated position and lowest in the
flexed position. The intraosseous circulation decreases
with elevation of the intra-articular pressure above the
systolic blood pressure, joint tamponade compressing
the periarticular tissues to such a degree that blood flow
reduces. In rabbits, the subchondral blood flow in the
femoral head begins to decline at an intracapsular
pressure of 40 cm of water. It is of note that joint
tamponade at a pressure below that of the arterial
circulation does not seriously impair the blood flow in
the femoral head.

22

Yet, absence of effective epiphyseal-

metaphyseal anastomoses, e.g. in the rat, puts the
femoral head at an increased risk. The blood supply of
the head is jeopardized in animals whose capsular vessels
are constrained by an intra-articular haemorrhage,

EXPERIMENTAL AVASCULAR OSTEONECROSIS

59

particularly if either traction or compression is con-
currently applied. The elevated pressure by itself
does not lead to an ischaemic catastrophe, rather it takes
a combination of perilous events to cause osteonecrosis.
As documented by a radioactive tracer microsphere test,
blood flow is reduced in the proximal femur of pigs with
a hip joint tamponade induced by an intra-articular
instillation of a dextran solution.

23

The reparative processes are, as a rule, absent

centrally and not well-developed peripherally in patients’
femoral heads many months after the beginning of
necrosis. By contrast, the necrotic debris in the rats’
femoral head is for the most part, or totally, replaced by
viable tissues within a few weeks of stripping the cervical
periosteum and cutting the ligamentum teres. It takes
just a few days for a flourishing fibrous tissue to spread
throughout the necrotic femoral head. This tissue, the
leading recovery signal, emanates mainly from the
inflamed and hyperplastic joint capsule. In addition to
the macrophages and osteoclasts, it delivers the
undifferentiated mesenchymal cells, i.e. progenitors of
fibroblasts, angioblasts, chondroblasts and osteoblasts.
Maturation of the fibrous tissue synchronizes with the
formation of new bone and cartilage, the latter to
a limited degree only. Repopulation of the intertrabecular
spaces by haematopoietic tissue takes place at this time.
Healing is complete by the 6th postoperative week in
some rats and by the 2nd postoperative month in
virtually all rats. New bone formation often results in
a shift from the normal spongy to a compact-like
construct. This bone’s composition is abnormal and its
mechanical properties, i.e.

stiffness, strength and

toughness, are inferior to those of mature bone. During
the modelling phase, everyday strains of load transfer
suffice to deform the epiphysis, lead to collapse of the
femoral head. Uniaxial compression loading tests
disclose deficient material properties of the mid and
late-stage postnecrotically healed cancellous bone. The
articular cartilage is involved in the remodelling process,
but it also degenerates consequent upon the altered
mechanical situation. This is linked to thickening and
condensation of the subchondral bone and, hence, to its
increased stiffness. The reduced yield strength, reduced
elastic modulus, and increased strain-to-failure attributes
of a healed infarcted bone, account for the subchondral
bone cave-in and the ensuing osteoarthritis-like disorder.
Epiphyseal-metaphyseal bony bridges bisect some rats’
physeal cartilage. The femoral head may be subtotally
absorbed in the most severe cases.

1,10,16,24

Radiographically graded, the measure of necrosis of

patients’ femoral heads ranges from mild (less than 15%
involvement) to severe (over 30% involvement). The
rats’ vascular-deprived femoral heads sustain total
necrosis in practically every case, expressing an
approximation (rather than equivalence) to the clinical
setting and its experimental model. This distinction, one

of several dissimilarities between man’s and animal’s
osteonecrosis, cannot be brushed aside off-hand. It is
contingent on dimensional, anatomical and patho-
physiological characteristics, species peculiarities, and
notably on the artificiality of the experimental setup. A
surgically induced sudden stoppage of both the arterial
and venous circulation in the femoral head of an
otherwise robust animal does not harmonize with our
perceptions of osteonecrosis in man.

25

The circulation in a dog’s necrotic femoral head is

re-established 3d4 months postoperatively. In com-
parison, the revascularization of a replanted avascular
spongy bone takes just 3 weeks. The expeditious
organization of the necrotic tissues subjacent to a
fracture site contrasts with the drawn-out ingrowth
of fibrous tissue into a circulation-deprived necrotic
bone. This highlights again the distinctive place assumed
by avascular osteonecrosis in the context of replace-
ment of dead by viable bone. Raising tissue oxygen
tension by exposure of rats to hyperbaric oxygen
facilitates the repair processes. It enhances osteoclast
and macrophage resorption of the debris and promo-
tes the proliferative as well as synthetic potential
of the fibroblasts, angioblasts and osteoblasts. Six weeks
postoperatively, less debris and more newly formed
bone are found in the necrotic femoral heads of rats
treated

with

hyperbaric

oxygen

than

in

those

untreated.

19

Providing a route for tissue penetration, bone grows

in and around a tract drilled from the greater trochanter
into avascular necrotic femoral heads. The blood
resupply advances more rapidly in drilled than in
untreated necrotic femoral heads. Transfer of a vas-
cularized bone graft into dogs’ necrotic femoral heads
promotes the growth of the terminal branches of the
medullary and periosteal arteries of the implant into
the necrotic bone.

26,27

Transplantation of blood vessels

augments reparative responses in a dog’s ischaemic
femoral head. Yet, the revascularization fails to fan out to
those peripheral regions of the head where osteogenesis
is enhanced at the site of the implanted vessels.

28

Because

of fibrous luminal obliteration, the effect of a drilling-
induced

decompression

is

shortlived.

Anticipating

prolongation of the decompression effects, resorbable
stents have been inserted in channels drilled into sheep’s
necrotic femoral heads. Growing along the channel from
the marrow and periosteum, sprouting vessels establish
anastomoses between the periosteal-diaphyseal-meta-
physeal and the epiphyseal-physeal vascular networks.

29

The femoral capital feeding vessels originate from the

medial and the lateral circumflex femoral arteries and
from vessels anastomosing with medullary arteries of the
neck. It is challenging to produce necrosis of a femoral
head in the face of a preserved marrow cavity and
obliterated physis. The blood flow in a femoral head of
a pig with a ligated deep femoral or lateral femoral

60

CURRENT ORTHOPAEDICS

Figure 4

An immature and mildly inflamed fibrous tissue dir-

ectly underlies the degenerated articular cartilage following seg-
mental resorption of the subchondral bone plate. Osteoclasts and
chondroclasts (arrows) abut the bone and cartilage. (Reproduced
from Pathology, Research and Practice with permission of Urban
& Fischer Verlag).

circumflex artery falls when the animal’s hip is held in the
frog-leg position. The blood flow declines to about 15%
of the baseline value in animals with dislocated hips and
ligated medial and lateral circumflex femoral arteries and
veins. The femoral head survives either the dislocation
or the ligation, but necrosis ensues by the 2nd to 4th
postoperative week in animals subjected to both
procedures. Results of quantitative assays substantiate
the emphasis given to the amounts of blood supplied to
the bone: a decrease below 20% of the baseline value
is a prerequisite for femoral capital necrosis to develop.
This fact explains why without an additional manipulation
osteonecrosis of the femoral head does not follow on
disruption of piglets’ medial femoral circumflex vessels.

30

ANIMAL MODELS OF PERTHES
DISEASE

The small amounts of blood that traverse a physis hardly
contribute to the nutrition of the femoral capital
epiphysis. In view of the coexistent effusion, transient
synovitis possibly plays a causative role in Perthes
disease. Raised intra-articular pressure is thought to
compress the intracapsular blood vessels to such a
degree that the ensuing ischaemia leads to epiphyseal
necrosis. The results of experimental modelling have not
corroborated this conjecture. The blood flow rate in the
femoral capital epiphysis of puppies whose intra-articular
pressure is raised to 50 mmHg is maintained within the
normal range. It takes a drastic pressure elevation to
values as high as 150 mmHg in order to disrupt the blood
circulation in the head.

31

Distension of the articular cavity

does not by itself result in epiphyseal ischaemia, despite
its tamponade effect. The femoral capital epiphysis is
deprived of blood, however, when articular cavity
distension concurs with joint compression. Obviously,
this holds true only when there is a persisting physeal
cartilage. An adult dog’s femoral head is nourished by the
capsular and medullary blood vessels such that one
vascular network is capable of vicariously substituting for
the other.

32

Necrosis of piglets’ femoral capital epiphyses is

outright on the 4th day after suture ligation of the
nourishing vessels. Areas of low signal intensity
located 3d8 weeks postoperatively on the T1 and T2-
weighted magnetic resonance images correspond to
chondronecrotic foci. Babyn and his associates’ paper is
one of but a few publications which explicitly brings up
chondrolysis as a complication, albeit a rare one, of
ischaemia-induced osteonecrosis.

24,33

The reparative process in the femoral epiphyses of

dogs

with

a

spontaneous

Perthes-like

disorder

terminates with collapse of the head. Osteoclasts and
macrophages cluster at the sites of bone resorption.
Enchondral ossification of the growth plate is disturbed

and arrested by necrotic foci in the metaphysis. Except
for small central foci, the epiphysis is revascularized at
the end of the reparative phase. As in the rat model
(Fig. 4), fibrous tissue violates the bone-cartilage junction
at the sites of resorption of the necrotic subchondral
bone, focally penetrating into the articular cartilage.
Acting in concert with the asymmetric growth and
mechanically

produced

collapse,

the

disorganized

enchondral ossification leads to flattening of the femoral
head.

34

Clogging the arteries of minipigs’ hindlimbs, injection

of glass microspheres causes proximal femoral ischaemia,
which is sometimes accompanied by focal necrosis of the
physeal cartilage and now and then of the metaphyseal
bone. Radiologically, the findings are reminiscent of the
features of Perthes disease in children. Metaphyseal
necrosis makes up part of a remote aftermath of
epiphyseal ischaemia. The focal breakdown of the growth
plate cartilage is a forerunner of the deformation of the
femoral head.

35

Epiphyseal modelling coexists with uni- or multifocal

bisection of the physis by osseous bridges in a quarter
of rats with vascular deprivation-induced necrosis of the
femoral head. The transphyseal osseous trabeculae,
which are produced within gaps emerging in the physis
after necrotic chondrocytes have been resorbed,
connect the epiphyseal and metaphyseal bone with one
another.

24

Under different circumstances, transphyseal

osseous bridges arise consequent on abnormal loading
of a remodeled femoral head: they bisect metacarpal,
metatarsal, and radial physes of calves housed on a clay
floor or a metal slat until weaning and, then, on a
concrete floor until the animals are 5-months old.
Eosinophilic, longitudinal streaks or horizontal patches,

EXPERIMENTAL AVASCULAR OSTEONECROSIS

61

Figure 5

Advanced modelling of the articular aspect of a rat’s

femoral head one month after disruption of all nutrient vessels.
Remnants of the degenerated articular cartilage (C) are covered
by a pannus (P). Where the subchondral bone has been resor-
bed, the residual cartilage overlays a vascularized fibrous tissue
(asterisks), containing undifferentiated mesenchymal cells (thin
arrows). The jagged aspect of the upper boundary of the osseous
trabeculae (thick arrows) is characteristic of the ongoing resorp-
tion of bone, fitting in with the presence of the osteoclasts. Note
also the optically empty osteocytic lacunae of the necrotic bone.
(Reproduced from Pathology, Research and Practice with per-
mission of Urban & Fischer Verlag).

which interrupt the physeal continuity, consist of
necrotic shreds of cartilage scattered along the plate
were found histologically. Residues of the necrotic
cartilage and hypertrophic chondrocytes are dispersed in
the metaphysis nearby the disarrayed physis. There is
an impressive resemblance to Perthes disease. A thin-
ned and broken-up cartilaginous plate underlies a re-
modelled epiphysis and overlies a modelled metaphysis.

36

Separated

from

each

other

by

an

unbroken

cartilaginous plate, the blood vessels of the epiphysis and
metaphysis make up independent circulations in adult
animals of species with a lifelong persisting physis. The
rat is a case in point. The circulation in murine femoral
heads comes to a total standstill following stripping of the
cervical periosteum and cutting the ligamentum teres.
Coagulation necrosis of the intertrabecular fibrous,
adipocytic and haematopoietic tissues is apparent on the
second day of ischaemia. Although the bone cells are also
injured early on, stainable osteocytic nuclei (in contrast
to the nuclei of other ischaemically damaged organs)
remain for some time after the lethal injury. It takes a few
days until the osteocytic lacunae of the dead bone begin
to appear ‘optically empty’. Thus, necrosis of the rat’s
circulation-deprived femoral head is morphologically first
obvious on the 5th postoperative day. The necrotic
intertrabecular debris, and subchondral and trabecular
bone are resorbed within 4 to 6 weeks of the operation
(Fig. 5). Concurrently, fibrogenesis, and appositional and
intramembranous osteogenesis lead to modelling of the

femoral head. This unfolding of the necrotic-to-
postnecrotic stage reduplicates the phases of Perthes
disease, albeit in an expedited time sequence, similar to
the time course of other repair events in small
laboratory animals. Clinicians are well aware that it does
not take long for fibrous tissue to permeate throughout
intertrabecular spaces of a patient’s necrotic bone, but
all-out revascularization and osseous restoration are
lengthy processes.

10

The alterations in the femoral head of spontaneously

hypertensive rats serve as a useful model for the study of
Perthes disease. Samples obtained from 9-week-old rats,
the age at which these animals are most prone to
osteonecrosis, reveal an obstruction of the lateral
epiphyseal arteries at the sites of their entry into the
femoral head. The vascular impediment precedes the
first signs of necrosis of the corresponding site in
the femoral head. As in other models, flattening of the
femoral head and shortening as well as widening of the
neck, best studied in 20-week-old rats, are secondary to
the reparative activities.

37

Dogs are sometimes afflicted

by a naturally occurring Perthes-like disease, which
apparently has a genetic background. The disorder is
inherited in Terriers. The transitorily disrupted venous
blood flow in the puppies’ femoral head yields a chain of
events, which with time triggers re-modelling terminating
in Perthes disease-like pathologic features.

38

Transient synovitis may contribute to the pathognesis

of Perthes disease. The effusion-related elevated intra-
articular pressure impedes the flow of blood in the
sub-synovial capsular vessels supplying the femoral
capital epiphysis. Raising the intra-articular pressure in
a pig’s hip to 150 mmHg for 10 h causes ischaemic
necrosis of the femoral epiphysis. Intriguingly at first
sight, raising the pressure to 200 mmHg by infusion of
autologous serum into the joint of a pig with talcum-
induced synovitis does not lead to epiphyseal necrosis.
This is in fact the expected outcome. The pressure in the
joint falls to 35 mmHg within 2 h, the inflamed joint
capsule stretching and the infusate oozing out of the joint
cavity. Replicating the clinical setting of a child with
a synovitic effusion, the results of this experimental set-
up cast doubts on the notion that articular tamponade of
the hip is a major causative player in Perthes disease.

39

TRAUMATIC OSTEONECROSIS

The pathogenesis of osteonecrosis is obvious in the
patient whose fractured bone or dislocated joint
interferes with the circulation, both the blood supply to
and the drainage from, say, the femoral head having been
stopped as all the periarticular vessels are torn. The
functionally effective anastomoses in an adult rabbit’s
femoral head, crossing from the metaphysis to the
epiphysis across the obliterated physis, minimize the

62

CURRENT ORTHOPAEDICS

consequence of disruption of the extraosseous blood
vessels. Posterior dislocation of a dog’s or a rabbit’s hip
hinders the circulation in the femoral circumflex arteries
and their metaphyseal and epiphyseal branches as well as
in the retinacular artery of the articular vascular circle.
Some arteries are disrupted within the torn articular
capsule and ligamentum teres. Although many vessels
survive the dislocation, the circulation is impeded by
arterial compression, traction and spasm. Fixed in an
anomalous position, the joint deformity of a dog’s
unreduced dislocated hip combines with reactive
inflammation and thrombotic occlusion or fibrotic
obliteration of the arteries to maintain or aggravate
the circulatory failure. The intraosseous blood flow may
be reduced to such a degree that ischaemic necrosis
of the femoral head ensues. The timely reduction of
the dislocated hip averts the vascular compression and
traction relieves the arterial spasm, and prevents or
alleviates the vascular obstruction. The restored
circulation

forestalls

the

development

of

osteo-

necrosis.

40

The limited circulation in the femoral heads of guinea

pigs with a fractured femoral neck is, to a degree, the
consequence of disrupted venous drainage. Laser
Doppler flowmetrical, microradiographical, densito-
metrical, and fluorescent label histological data highlight
a delayed revascularization and re-modelling, which
reaches a climax 5d7 weeks after the ‘traumatic’ event.
Bereft of their mechanical strength, the thinned, re-
modelled subchondral plate and osseous trabeculae
equally contribute to the collapse of the femoral head. As
far as the extent of the necrosis in the femoral head is
concerned, the histological and mechanical scenes in this
model approximate the clinical spectrum in patients with
a femoral neck fracture.

41

Although this review is focused on experimental

osteonecrosis, Brown and coworkers’ conceptions
should be kept in mind. The biomechanical integrity of
the subchondral bone overlying the weak cancellous
bone of a necrotic femoral head is preserved during
the early stages. In a three-dimensional finite element
analysis, the authors have verified the proposal that
a biomechanically sufficient subchondral bone provides
ample stress protection for the underlying weakened
bone. Stress levels in the cancellous bone of a necrotic
femoral head are around 70% of the stress levels in the
subchondral bone. With the passage of time, the subchon-
dral bone deteriorates in the wake of the ongoing
re-modelling. Yet, the principal stress distribution
in this weak subchondral bone hardly differs from
that in the subchondral plate of a normal femoral head.
The ratio of the stress to strength reflects the tendency
of structural failure to be critically higher in the
cancellous than in the subchondral bone. These data
convey a clinically salient principle. Even the normal,
healthy subchondral bone plate provides just a modest

stress protection for an underlying, diseased and
weakened cancellous bone. At the same time, the stress
does increase in the subchondral bone overlying a fragile
cancellous bone. Therefore, onset of the collapse of a
necrotic femoral head is dominated by the extent and
intensity of the structural degradation of the cancellous
bone within the main territory of the infarcted femoral
head rather than by the degree of the structural
degradation of the subchondral bone plate.

42

CORTICOSTEROID
ADMINISTRATION-RELATED
OSTEONECROSIS

The pathways leading to osteonecrosis in the cortico-
steroid-treated animals are manifold and multifold.
It has been gleaned on gauging the intraosseous pressure
that vascular lesions antedate intraosseous hyper-
tension of corticosteroid-treated rabbits by close to 3
weeks. The arterioles, primarily the subchondral ones,
are plugged by fibrin-platelet thrombi, which are loaded
with lipids emanating from the steroid-induced fatty liver.
The steroid treatment-related hypertrophic lipocytes
compress the venules and small veins, resulting in
capillary stasis. The blood flow in the veins is slowed
down. Overall, the blood flow is reduced by nearly one
third after a 10-week steroid regimen. A falling blood
supply of this magnitude does not by itself result in
osteonecrosis, but it contributes to cell death incidental
to other factors. Osteonecrosis is first evident in rabbits’
femoral heads by the 2nd week of methylprednisolone
medication and is widely spread in both the proximal
and the distal segments of the femur by the 18th week.
Developing in the fragile porotic bones, microfractures
compress subchondral vessels, worsening the already
precarious circulation. The blood supply increases
following core decompression, normalizing in about
1 month. Ingrowth of sprouting granulation tissue ushers
in the revascularization of the necrotic bone. Myo-
fibroblasts, an intrinsic constituent of all granulation
tissues, produce an endothelial cell growth factor which
stimulates angiogenesis. Albeit not influencing endo-
thelial cells directly, cortiocosteroids inhibit capillary
proliferation by suppressing myofibroblastic functions.
It is premature to speculate at this time to what
extent, if any, steroid-induced impeded angiogenesis
impacts on events occurring during the early, the late
or both phases of osteonecrosis.

43,44

The haematopoietic tissue undergoes ischaemic

necrosis in rabbits’ metaphyseal bones during the early
stage of horse serum-induced hypersensitivity vasculitis,
at a time that the circulation is arrested in the inflamed
terminal arterioles. Acting in concert with the untoward
effects of the corticosteroids, the hypersensitivity
vasculitis leads to medullary haemorrhages and, often, to

EXPERIMENTAL AVASCULAR OSTEONECROSIS

63

death of the trabecular bone. Histologically, the number
of discernible arterioles, principally in the subchondral
zone, is smaller in the medulla comprising steroid-
induced hypertrophic adipocytes than in an untreated
rabbit’s medulla containing normally sized fat cells. This
combination of circumstances is responsible for the,
maybe at first sight unforeseen, high risk of necrosis
of the femoral head of steroid-treated animals with
hypersensitivity vasculitis.

45

PHYSICAL INJURY-INDUCED
OSTEONECROSIS

Traumatic osteonecrosis, the result of high energy
physical injuries, has been discussed above. Drastic
alterations of the ambient temperature, in either
direction lethally injure the soft and hard tissues. Low
temperatures dissipated by a liquid nitrogen cryoprobe
inserted into rabbits’ medullary cavity bring about bone
infarction secondary to freezing-induced thrombotic
occlusion of the small vessels. Not icy temperatures by
themselves, but ischaemia is the paramount trigger of
necrosis after cryo-surgery. Periosteal osteogenesis
dominates the early recovery phase. In the absence of
a periosteum wrapping the articulating portion of the
bone, all repair processes are restricted to the deep
compartment of the femoral head. This peculiarity
is clinically critical. Generous periosteal new bone
formation augments the strength of a necrotic bone at
the early healing stages. Such a strengthening mechanism
is not operative in the hemispheric part of the necrotic
femoral head. In the dogs’ lateral femoral condyle,
cryotreatment of a cortiococancellous core destroys
a bone volume surpassing almost three times the
dimension of the surgically prepared cavity, the necrotic
area extending between 2.5 and 14 mm beyond the
cavity walls. Abnormal osteogenesis persists for nearly
7 weeks in cold-injured bones. Regeneration of the
osseous tissue is initially scanty, but the replacement of
the dead by viable bone then catches up, healing being
complete within 24 weeks of the freezing event.

46

That hyperthermia is deadly to cells is well known. The

threshold at which the thermal stress consistently kills
the osteocytes is debatable. Raising the temperature in
a dog’s tibial medulla to 42.53C for 60 min causes
necrosis of the bone to a distance of 5 mm from a heated
stainless steel nail. Inasmuch as the bone recovers within
just 12 weeks, hyperthemia, at least under these
circumstances, is but mildly harmful to the osseous
environment.

47

Osteoradionecrosis, complicating therapeutic irra-

diation, was recognized by the pioneers of radiobiology.
Osteocytes are lethally injured by X-ray doses which
do not kill other musculoskeletal cells. Three to
15 months after a 50-G irradiation of adult rabbits’

stifles, the subchondral bone is necrotic, but the articular
cartilage is ‘healthy’ by light microscopy (no degenera-
tive lesions), scanning electron microscopy (normal
collagen fibers pattern), and autoradiography (active
RNA synthesis).

48

DYSBARIC OSTEONECROSIS

Dysbaric osteonecrosis is the outcome of ill-suited and
ill-timed exposures to a hyperbaric environment. It has
been reproduced in sheep exposed, during a 2 month
period, to twelve, 24-h long sessions of breathing
compressed air (at 2.6d2.9 atmospheres absolute). The
blood flow through femoral heads of rabbits exposed for
4 hours to compressed air (at 3 atmospheres absolute) is
notably lower than through the femoral heads of animals
breathing air at a pressure of one atmosphere. The
reduced blood flow is not solely the result of the high
partial oxygen pressure. The blood flow through a guinea
pig’s femoral head falls after a 90-min exposure to air at
a pressure of 0.5 MPa followed by a stimulated ascent
rate of 0.03 MPa/min.

The relationship between the inadequate decom-

pression of the rabbits which have been exposed to
hyperbaric air and the reduced blood flow in their
femoral heads indicates the pivotal role played by
ischaemia in the causation of the osteonecrosis under
these circumstances. Injuring osteocytes irreversibly,
active oxygen species apparently contribute to dysbaric
osteonecrosis. In hyperbaric air-exposed swine, deeper
or repetitive ‘dives’, a low oxygen concentration, and
a rapid, indiscriminate decompression lead to a maximal
consumption of the haemostatic factors. With alleviation
of a swine’s diving profile, the platelets and fibrinogen are
less depleted. The destruction of endothelial cells and
the thickening of the arterial intima ensue under
conditions that exhaust fibrinogen and platelets. The
vascular lesions express reparative processes; their
extent peaks after discontinuation of the stimulated
diving sessions. The scope of gas bubble formation and
the degree of platelet and fibrinogen consumption
parallel increasing ‘diving’ depths. The vessels are
obstructed from within by the gas bubbles, clumped
erythrocytes, and coalesced lipids and are compressed
from without by extravascular gas bubbles. By injuring
the endothelial cells, the bubbles activate aggregation of
the platelets and precipitation of the fibrin. Reactions at
blood-bubble interfaces trigger the coagulation cascade,
boost release of vasoactive agents, and activate gas-
induced osmosis. The bubbles are indirectly responsible
for the arterial narrowing, as activated platelets release
growth factors which stimulate proliferation of the
myointimal cells. In hindering smooth currents, rheologic
permutations exert an ancillary role. Gas super-
saturation of the fatty marrow, tissue rigidity-dependent

64

CURRENT ORTHOPAEDICS

sensitivity to the extravascular gas pressure, relatively
poor vascularization, and presence of nucleation and
bubble formation-promoting substances act concur-
rently or sequentially in adding towards the ischaemic
insult.

The extent to which the body’s build impacts on the

disorder is remarkable. Necrosis of the femoral and tibial
epiphyses evolves in a third of obese mice but in less than
one tenth of lean animals subjected to an identical
hyperbaric ambience for the same length of time.
Besides, the latent period in the obese mice is shorter
than in the lean animals. With involvement of the juxta-
articular bone, the death of osteocytes in the sub-
chondral zone, marrow fibrosis, reduplicated tidemark,
and thinned joint cartilage suggest an early onset of
osteoarthritis. Finally, it is recalled that dysbaric osteo-
necrosis and decompression sickness are distinct from
one another, though they share several aetiologic factors
and pathogenic pathways.

49d51

CONCLUDING REMARKS

Because of space limitation, the many insights gained by
experimenters during the preceding two decades in the
course of investigations of osteonecrosis in animals has
not been discoursed exhaustively. Shwartzman reaction,
chemically, and

metabolically induced

variants of

osteonecrosis have not been cited. Immune reaction-
mediated osteonecrosis is alluded to in passing only.
Fundamentally, aseptic osteonecrosis is the outcome of
an imbalance between the demand and supply of
nutrients, primarily oxygen, as a result of an impaired
circulation due to either a local or a systemic
disturbance. Clinically, for many patientsealbeit cur-
rently for but a minority of themeovert risk factors cannot
be uncovered. The epithet ‘idiopathic osteonecrosis’
is assigned to this group of patients.

Except for its idiopathic version, osteonecrosis

complicates a variety of diseases and syndromes. Cruess
has well appreciated its complex nature. Impediments of
the micro and macrovasculature from the inside, e.g.
microembolic events, and from the outside, e.g. adipo-
cytic hypertrophy, as well as lipid-incurred osteocytic
death, alone or in combination, play aetiological and
pathogenic roles. The critique of experimentations
designed to analogize healing of ‘infarcted’ bones with
vascularization of grafted bones is correct. Deductions
based on the findings in models are burdened by the fact
that the interventions, be they surgical, chemical, physical
or otherwise, are conducted on healthy animals, the
muscles and bones of which are of excellent quality.
On the other hand, osteonecrosis at the bedside is met
with, more often than not, in patients with one or
another musculoskeletal or systemic disease, say, in
a

corticosteroid-treated

patients

with

an

organ

transplant or systemic lupus erythematosus.

Hsieh et al. rightfully reason that ‘to understand the

pathophysiology of a disease, one must reproduce it in an
animal model where it can be monitored throughout its
course’. Investigators have generally not achieved
adequate models of ischaemic osteonecrosis.

52

Those

who have aspired to ascertain generic aetiological and
pathogenic principles common to all categories of
osteonecrosis have failed in their quest. Not all research
yields data which are helpful in everyday medical practice.

Morphologic analysis is fraught with difficulties. In

a non-osseous tissue, necrotic cells display karyopyknosis,
karyorrhexis, and karyolysis within hours to day or so of
a deadly insult. This does not hold true for the bone.
Well-stained nuclei are encountered for as long as 16
weeks after osteocytes have been lethally damaged. It is
customary to adhere to the guidelines which state that
loss of stainable nuclei (karyolysis) and cytoplasmic
structures (cytolysis), karyopyknosis and karyorrhexis of
fibrous, fatty, vascular, and haematopoietic tissues of
the intertrabecular space define necrosis of the marrow.
The presence of optically empty lacunae of the vast
majority of the osteocytes speaks for early stage
osteonecrosis,

while

appositionally

formed

bone

alongside the necrotic bone evidences advanced stage
osteonecrosis.

53

Kenzora and Glimcher argue that non-traumatic

osteonecrosis is aetiologically and pathogenically multi-
farious in nature. The functionally distressed osteocytes,
the result of any of a number of diverse factors,
are increasingly damaged with progression of an under-
lying disease or incidentally operative agents until cellular
regression gets to the point at which equilibrium is no
longer maintainable. For example, constituting the pre-
ponderant ultimate stress, the corticosteroids add their
noxious effects to one disease or another by augmenting
the bone marrow fat stores, embolizing fat droplets,
raising the intraosseous pressure, or by directly exerting
a toxic action on the bone cells. Restriction of the
necrotic zone to a certain skeletal site is accounted for
where a causative mechanism is operative locally, say,
high-energy trauma or thromboembolic occlusion of
an artery.

Where causative mechanisms operate

systemically, explanations for the restriction of the
necrotic zone to particular skeletal sites are intricate,
how much more so in cases of idiopathic osteonecrosis.
No

conjectures

fully

rationalize

susceptibility

of

particular skeletal sites, such as the femoral head,
femoral condyles, and humeral head. Still, the significance
of an unfavorable circulatory situation should be
reiterated, e.g. all major vessels entering the bone from
one direction. These reviewers are unaware of an
experimental approach which specifically addresses this
subject matter. Based on positron emission tomographic
findings in young volunteers, it is evident that any
regional variations in the blood volume of the bone
marrow correlate with the different incidence of

EXPERIMENTAL AVASCULAR OSTEONECROSIS

65

ischaemic

osteonecrosis

in

the

several

skeletal

regions.

54,55

That some modernistic research is dedicated to bone

cell apoptosis is expected, up-to-date medical research
being committed to this phenomenon. By definition,
necrosis is an accidental cell death while apoptosis
a programmed cell death. The subject of osteoblastic and
osteocytic apoptosis has not been reviewed because it
is currently recommended to ‘use the term necrosis
to describe findings comprising dead cells in histological
sections, regardless of the pathway by which the cells
died’.

56

Applying the terminal deoxynucleotidyl trans-

ferase-mediated

deoxyuridine

triphosphate

biotin

nick end labelling technique to display apoptosis-
associated fragmented DNA, Kabata et al. have located
apoptotic cells in specimens obtained from rabbits
with early stage corticosteroidinduced osteonecrosis.

57

We look forward to learning more about break-
throughs in this fast advancing field in the foreseeable
future.

REFERENCES

1. Mont M, Hungerford D. Non-traumatic avascular necrosis of the

femoral head. J Bone Joint Surg 1995; 77A: 459d474.

2. Winet H, Hsieh A, Bao J Y. Approaches to study of ischemia in

bone. J Biomed Mater Res 1998; 43: 410d421.

3. Bullough P G, Kambolis C P, Marcove R C, Jaffe H L. Bone

infarctions not associated with caisson disease. J Bone Joint Surg
1965; 47A: 477d491.

4. Glowacki J. Angiogenesis in fracture repair. Clin Orthop 1998; 355

Suppl: 82d89.

5. Takaoka K, Yoshioka T, Hosoya T, Ono K, Takase T. The repair

process in experimentally induced avascular necrosis of the
femoral head in dogs. Arch Orthop Trauma Surg 1981; 99:
109d115.

6. Kenzora J E. The multifactorial etiology of idiopathic osteonecrosis

associated with corticosteroid administration. J Rheumatol 1983;
10: 48d58.

7. Jones J P Jr. Etiology and pathogenesis of osteonecrosis. Sem

Arthroplasty 1991; 2: 160d168.

8. Solomon L. Mechanisms of idiopathic osteonecrosis. Orthop Clin

North Am 1985; 16: 655d657.

9. Nelson G E, Kelly P J, Peterson F A, Janes J M. Blood supply of the

human tibia. J Bone Joint Surg 1960; 42A: 625d636.

10. Levin D, Norman D, Zinman C, Misselevich I, Reis N R, Boss J H.

Osteoarthritis-like disorder in rats with vascular deprivation-
induced necrosis of the femoral head. Pathol Res Prac 1999; 195:
637d642.

11. Beppu M, Takahashi F, Tsai T M, Ogden L, Sharp J B Jr.

Experimental replan-tation of canine forelimbs after 78.5 hours of
anoxia. Plast Reconstr Surg 1989; 84: 642d648.

12. Launder W J, Hungerford D S, Jones L C. Hemodynamics of the

femoral head. J Bone Joint Surg 1981; 63A: 442d448.

13. Rhinelander F W, Nelson C L, Stewart R D, Stewart C L.

Experimental reaming of the proximal femur and acrylic cement
implantation: Vascular and histologic effects. Clin Orthop 1979;
141: 74d89.

14. Klein L R, Dollinger B, Goldberg V M, Zika J M, Powell A E, Heiple

K G. Effects on bone of vascular interruption. Turnover and
morphology in isotope-prelabelled rats. Acta Orthop Scand 1985:
56: 47d51.

15. Whiteside L A, Lange D R, Capello W R, Fraser B. The effects of

surgical procedures on the blood supply to the femoral head.
J Bone Joint Surg 1983; 65A: 1127d1133.

16. Norman D, Reis D, Zinman C, Misselevich I, Boss J H. Vascular

deprivation-induced necrosis of the femoral head of the rat. An
experimental model of avascular osteonecrosis in the skeletally
immature individual or Legg-Perthes disease. Int J Exp Pathol 1998;
79: 173d181.

17. Zweier J L, Broderick R, Kuppusamy P, Thompson-Gorman S,

Lutty G A. Determination of the mechanism of free radical
generation in human aortic endothelial cells exposed to anoxia and
reoxygenation. J Biol Chem 1994; 269: 24156d241562.

18. Brown T D, Way M E, Ferguson A B Jr. Mechanical characteristics

of bone in femoral capital aseptic necrosis. Clin Orthop 1981; 156:
240d247.

19. Levin D, Norman D, Zinman C, Rubinstein L, Sabo E, Misselevich I,

Reis D, Boss J H. Treatment of experimental avascular necrosis of
the femoral head with hyperbaric oxygen in rats: Histological
evaluation of the femoral heads during the early phase of the
reparative process. Exp Molec Path 1999; 67: 99d108.

20. Divakov M G. Revascularization of avascular spongy bone and head

of the femur in transplantation of vascular bundle (An experimental
and clinical study). Acta Chir Plast 1991; 33: 114d125.

21. Calvert P T, Kernohan J G, Sayers D C, Catterall A. Effects of

vascular occlusion on the femoral head in growing rabbits. Acta
Orthop Scand 1984; 55: 526d530.

22. Krebs B, Nue Moller B, Falstie-Jensen S, Love-Jepsen F, Taagehoj

Jensen F. Scintimetry of hip joint tamponade in dogs. Acta Orthop
Scand 1986; 57: 111d114.

23. Downey D J, Simkin P A, Lanzer W L, Matsen F A. Hydraulic

resistance: A measure of vascular outflow obstruction in
osteonecrosis. J Orthop Res 1988; 6: 272d278.

24. Peskin B, Shupak A, Misselevich I, Zinman C, Levin D, Jacob Z, Reis

D, Boss J H. Transphyseal osseous bridges in experimental
osteonecrosis of the femoral head of the rat: Development of bony
trabeculae connecting the epiphysis with the metaphysis through
localized areas of lost physeal cartilage. J Pediatr Orthop, in press.

25. Steinberg M E, Hayken G D, Steinberg D R. A quantitative system

for staging avascular necrosis. J Bone Joint Surg 1995; 77B: 34d41.

26. Dahners L E, Hillsgrove. The effects of drilling on revascularization

and new bone formation in canine femoral heads with avascular
necrosis: An initial study. J Orthop Trauma 1989; 3: 309d312.

27. Uchida Y, Sugioka Y. Effects of vascularized bone graft on

surrounding necrotic bone: an experimental study. J Reconstr
Microsurg 1990; 6: 101d107.

28. Gartsman G M, Weiland A J, Moore J R, Randolph M A. Blood

vessel implantation into ischemic bone. J Reconstr Microsurg 1985;
1: 215d222.

29. Simank H G, Graf J, Kerber A; Wiedmaier S. Long-term effects of

core decompression by drilling. Demonstration of bone healing
and vessel ingrowth in an animal study. Acta Anat 1997; 158:
185d191.

30. Jaramillo D, Villegas-Medina O L, Doty D K, Dwek J R, Ransil B J,

Mulkern R V, Shapiro F. Gadolinium-enhanced M R imaging
demonstrates abduction-caused hip ischemia and its reversal in
piglets. A J R Am J Roentgenol 1996; 166: 879d887.

31. Borgsmiller W K, Whiteside L A, Goldsand E M, Lange D R. The

effect of hydrostatic pressure in the hip joint on proximal femoral
epiphyseal and metaphyseal blood flow. Proc Orthop Res Soc
1980; 5: 23.

32. Nakamura T, Matsumoto T, Nishino M, Tomita K, Kadoya M. Early

magnetic resonance imaging and histologic findings in a model of
femoral head necrosis. Clin Orthop 1997; 334: 68d72.

33. Babyn P S, Kim H K, Gahunia H K, Lemaire C, Salter R B, Fomasier

V, Pritzker K P J. MRI of the cartilaginous epiphysis of femoral head
in the piglet hip after ischemic damage. Magn Reson Imaging 1998;
8: 717d723.

66

CURRENT ORTHOPAEDICS

34. Mickelson M R, McCurnin D M, Awbrey B J, Maynard J A, Martin

R K. Legg-Calve-Perthese disease in dogs: a comparison to human
Legg-Calve-Perthes disease. Clin Orthop 1981; 157: 287d300.

35. Smith S R, Malcolm A J, Gregg P J. Metaphyseal changes in Perthes’

disease: An experimental study. Br J Exp Pathol 1982; 63: 633d638.

36. White S L, Rowland G N, Whitlock R H. Radiographic,

macroscopic, and microscopic changes in growth plates of calves
raised on hard flooring. Am J Vet Res 1984; 45: 633d639.

37. Hirano T, Iwasaki K, Yamane Y. Osteonecrosis of the femoral head

of growing, spontaneously hypertensive rats. Acta Orthop Scand
1988; 59: 530d535.

38. Vasseur P B, Foley P, Stevenson S, Heitter D. Mode of inheritance

of Perthes’ disease in Manchester terriers. Clin Orthop 1989; 244:
281d292.

39. Gershuni D H, Hargens A R, Lee Y F, Greenberg E N. Zapf R,

Akeson W H. The questionable significance of hip joint tamponade
in producing osteonecrosis in Legg-Calve-Perthes syndrome.
J Pediatr Orthop 1983; 3: 280d286.

40. Shim S S. Circulatory and vascular changes in the hip following

traumatic hip dislocation. Clin Orthop 1979; 140: 255d261.

41. Swiontkowski M F, Tepic S, Rahn B A, Cordey J, Perren S M. The

effect of fracture on femoral head blood flow. Osteonecrosis and
revascularization studied in miniature swine. Acta Orthop Scand
1993; 64: 196d202.

42. Brown T D, Baker K J, Brand R A. Structural consequences of

subchondral bone involvement in segmental osteonecrosis of the
femoral head. J Orthop Res 1992; 10: 79d87.

43. Wang G J, Dughman S S, Reger S I, Stamp W G. The effect of core

decompression on femoral head blood flow in steroid-induced
avascular necrosis of the femoral head. J Bone Joint Surg Am 1985;
67A: 121d124.

44. Bouteiller G, Arlet J, Blasco A, Vigoni F, Elefterion A. Is

osteonecrosis of the femoral head avascular? Bone blood flow
measurements after long-term treatment with corticosteroids.
Metab Bone Dis Relat Res 1983; 4: 313d318.

45. Matsui M, Saito S, Ohzono K, Sugano N, Saito M, Takaoka K, Ono

K. Experimental steroid-induced osteonecrosis in adult rabbits
with hypersensitivity vasculitis. Clin Orthop 1992; 277: 61d72.

46. Keijser L C, Schreuder H W, Buma P, Weinans H, Veth R P.

Cryosurgery in long bones; an experimental study of necrosis and
revitalization in rabbits. Arch Orthop Trauma Surg 1999; 119:
440d444.

47. Ikenaga M, Ohura K, Kotoura Y, Yamamuro T, Nakamura T, Oka

M, Hiraoka M, Abe M. Hyperthermic treatment of canine tibia
through RF inductive heating of an intramedullary nail: A new
experimental approach to hyperthermia for metastatic bone
tumours. Int J Hyperthermia 1994; 10: 507d516.

48. Takahashi S, Sugimoto M, Kotoura Y, Oka M, Sasai K, Abe M,

Yamamuro T. Long-lasting tolerance of articular cartilage after
experimental intraoperative radiation in rabbits. Clin Orthop 1992;
275: 300d305.

49. Davis T R, Holloway I T, Pooley J. Effect of exposure to

compressed air and elevated oxygen levels on bone blood flow in
the rabbit. Undersea Biomed Res 1990; 17: 201d211.

50. Jones G R. Dysbaric osteonecrosis (caisson disease of bone): Are

active oxygen species and the endocrine system responsible, and
can control of the production of free radicals and their reaction
products confer protection? Free Radic Res Commun 1987; 4:
139d147.

51. Cruess

R

L.

Osteonecrosis

of

bone.

Current

concepts

as to etiology and pathogenesis. Clin Orthop 1986; 208:
30d39.

52. Hsieh A S, Winet H, Bao J Y, Stevanovic M. Model for intravital

microscopic evaluation of the effect of arterial occlusion-caused
ischemia in bone. Ann Biomed Eng 1999; 27: 508d516.

53. Sakai T, Sugano N, Tsuji T, Miyazawa T, Nakamura N, Haraguchi K,

Ochi T, Ohzono K. Contrast-enhanced magnetic resonance
imaging in a nontraumatic rabbit osteonecrosis model. J Orthop
Res 1999: 17: 784d792.

54. Kenzora J E, Glimcher M J. Accumulative cell stress: The

multifactorial etiology of idiopathic osteonecrosis. Orthop Clin
North Am 1985; 16: 669d679.

55. Iida S, Harada Y, Ikenoue S, Moriya H. Measurement of bone

marrow blood volume in the knee by positron emission
tomography. J Orthop Sci 1999; 4: 216d222.

56. Levin S, Bucci T J, Cohen S M, Fix A S, Hardisty J F, LeGrand E K,

Maronpot R R, Trump B F. The nomenclature of cell
death: Recommendations of an ad hoc Committee of the Society
of

Toxicologic

Pathologists.

Toxicol

Pathol

1999;

27:

484d490.

57. Kabata T, Kubo T, Matsumoto T, Nishino M, Tomita K, Katsuda S,

Horii T, Uto N, Kitajima I. Apoptotic cell death in steroid induced
osteonecrosis: An experimental study in rabbits. J Rheumatol 2000;
27: 2166d2171.

EXPERIMENTAL AVASCULAR OSTEONECROSIS

67