Rheum Dis Clin N Am

34 (2008) 531–559

Etiopathogenesis of Osteoarthritis
Kenneth D. Brandt, MDa,b,*, Paul Dieppe, MDc,
Eric L. Radin, MDd
a
Kansas University Medical Center, 5755 Windsor Drive,
Fairway, Kansas City, KS 66205, USA
b
Indiana University School of Medicine, 1100 West Michigan, Indianapolis, IN 64202, USA
c
Nuffield Orthopaedic Centre, Nuffield Department of Orthopaedic Surgery,
Windmill Road, Headington, Oxford, OX3 7LD, UK
d
Tufts University School of Medicine, 6 School St., PO Box 561, Marion, MA 02738, USA

Current definitions of osteoarthritis

At a 1986 workshop on the etiopathogenesis of osteoarthritis (OA) spon-
sored by the National Institute of Arthritis, Diabetes, Digestive, and Kidney
Diseases, the National Institute on Aging, the American Academy of Ortho-
paedic Surgeons, the National Arthritis Advisory Board and the Arthritis
Foundation, OA was defined comprehensively [1]:
Clinically, the disease is characterized by joint pain, tenderness, limitation
of movement, crepitus, occasional effusion, and variable degrees of local
inflammation, but without systemic effects. Pathologically, the disease is
characterized by irregularly distributed loss of cartilage more frequently
in areas of increased load, sclerosis of subchondral bone, subchondral
cysts, marginal osteophytes, increased metaphyseal blood flow, and vari-
able synovial inflammation. Histologically, the disease is characterized
early by fragmentation of the cartilage surface, cloning of chondrocytes,
vertical clefts in the cartilage, variable crystal deposition, remodeling, and
eventual violation of the tidemark by blood vessels. It is also characterized
by evidence of repair, particularly in osteophytes, and later by total loss
of cartilage, sclerosis, and focal osteonecrosis of the subchondral bone. Bio-
mechanically, the disease is characterized by alteration of the tensile, com-
pressive and shear properties and hydraulic permeability of the cartilage,

Some of the thoughts expressed in this article are contained in a chapter, ‘‘Neuromuscular
Aspects of Osteoarthritis,’’ written by KDB, that is to be published by Wiley-Blackwell in the
book, Osteoarthritis Pain, edited by D.T. Felson and H-G. Schaible.
* Corresponding author. Kansas University Medical Center, 5755 Windsor Drive,
Fairway, Kansas City, KS 66205.
E-mail address: [email protected] (K.D. Brandt).

0889-857X/08/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.rdc.2008.05.011 rheumatic.theclinics.com
532 BRANDT et al

increased water, and excessive swelling. These cartilage changes are accom-
panied by increased stiffness of the subchondral bone. Biochemically, the
disease is characterized by reduction in the proteoglycan concentration,
possible alterations in the size and aggregation of proteoglycans, alteration
in collagen fibril size and weave, and increased synthesis and degradation of
matrix macromolecules.
No more succinctly, 9 years later, at a workshop sponsored by the American
Academy of Orthopaedic Surgeons, the National Institute of Arthritis, Mus-
culoskeletal, and Skin Diseases, the National Institute on Aging, the Arthritis
Foundation, and the Orthopaedic Research and Education Foundation, OA
was redefined as follows [2]:

Osteoarthritis is a group of overlapping distinct diseases which may have

different etiologies, but with similar biologic, morphologic, and clinical out-
comes. The disease processes not only affect the articular cartilage, but
involve the entire joint, including the subchondral bone, ligaments, capsule,
synovial membrane, and periarticular muscles. Ultimately, the articular car-
tilage degenerates with fibrillation, fissures, ulceration, and full thickness loss
of the joint surface. OA diseases are a result of both mechanical and biologic
events that destabilize the normal coupling of degradation and synthesis of
articular cartilage of chondrocytes and extracellular matrix, and subchondral
bone. Although they may be initiated by multiple factors, including genetic,
developmental, metabolic, and traumatic, OA tissues involve all of the tissues
of the diarthrodial joint. Ultimately, OA diseases are manifested by morpho-
logic, biochemical, molecular, and biomechanical changes of both cells and
matrix which lead to a softening, fibrillation, ulceration, loss of articular
cartilage, sclerosis and eburnation of subchondral bone, osteophytes, and
subchondral cysts. When clinically evident, OA diseases are characterized
by joint pain, tenderness, limitation of movement, crepitus, occasional effu-
sion, and variable degrees of inflammation without systemic effects.
These inclusive definitionsdthe latter is the most current in the fieldd
offer something for everyone but are not helpful in understanding the
etiopathogenesis of OA. In both definitions the emphasis is on joint damage
in general and on the loss of articular cartilage in particular. Both fail to
recognize that OA reflects a process in the joint that is attempting to contain
damage that has been caused by a local mechanical problem.
Furthermore, neither of these definitions distinguishes between ‘‘garden
variety’’ OA, (ie, the common presentation of OA) and OA caused by rare
systemic diseases that have very different clinical presentations (eg, ochrono-
sis, the chondroepiphyseal dysplasia that results from a point mutation in
a gene that codes for type II procollagen [3], Gaucher’s disease, and other con-
ditions). Indeed, garden variety OA itself is not a homogeneous entity but may
arise as a result of a congenital or developmental anatomic abnormality of the
joint, obesity, trauma, ligamentous instability or micro-incoordination. The
long-held distinction between primary and secondary OA is not meaningful:
OA always is secondary to something.
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 533

Although the older of these definitions of OA refers to biomechanical

changes in articular cartilage and bone, it says nothing about abnormal
mechanics of the joint. The more recent definition considers OA to be the
‘‘result of mechanical and biologic events that destabilize’’ the normal met-
abolic equilibrium of cartilage and subchondral bone; however, like its pre-
decessor, this definition does not mention joint biomechanics. Box 1 lists
a number of facts about OA that are not apparent in the current definitions.

A current view of the etiopathogenesis of osteoarthritis

In the definitions given in the previous section, the emphasis is on joint
damage in general and on loss of articular cartilage in particular. In contrast,
the authors view OA as a process that is attempting to contain a mechanical
problem in the joint. OA is best defined as failed repair of damage that has
been caused by excessive mechanical stress (defined as force/unit area) on joint
tissues. Because the body’s innate mechanisms for repairing the damaged
tissues cannot be effective in the face of the underlying mechanical abnorma-
lity, they cannot solve the problem of OA; that is, remodeling of subchondral
bone may reduce the excessive stress and contain the mechanical abnormality
but may result in joint pain.

Box 1. Key points that are not recognized by the current

definitions of osteoarthritis
The clinical presentation of osteoarthritis
1. There are multiple causes of osteoarthritis.
2. The most common clinical presentation (ie, garden-variety
osteoarthritis) can result from a wide variety of insults
to the joint.
Etiopathogenesis
1. Osteoarthritis is initiated by a mechanical insult to the joint.
2. Osteoarthritis is a manifestation of attempts to heal the joint
and ameliorate the abnormal biomechanics.
3. The osteoarthritis process may cause joint pain but often is
successful, leading to a stable, painless joint.
Radiology of osteoarthritis
1. Radiographic changes of osteoarthritis are extremely common
in the population.
2. Many people who have severe radiographic changes of
osteoarthritis are asymptomatic.
3. Radiographic progression of osteoarthritis usually is slow and
may cease completely for many years.
534 BRANDT et al

Because OA represents the failure of an organ (the synovial joint), any of

the tissues of that organ may be the first to fail. It would be expected, there-
fore, that there are many causes of OAdand there are. OA has no common
pathophysiologic pathway but only a final common end stage [4]. In OA, in-
flammatory changes are secondary and are caused by particulate and soluble
breakdown products of cartilage and bone. OA should not be considered
a degenerative joint disease so long as the cells of the cartilage and bone
are normal and, if the high levels of intra-articular stress are reduced, can
restore the damaged tissue to normal [5]. Furthermore, apart from the inac-
curacy of the term, consider the pessimism, sense of futility, and nihilism
that are engendered in a patient who has painful OA when her physician
tells her she has degenerative joint disease.
The common mechanical factor underlying OA is a pathologic increase in
intra-articular stress, which can result either from a decrease in the load-
bearing area on the joint surface or from a quantitative increase or qualitative
aberration in joint loading (ie, repetitive impulsive loading) [6]. Excessive
loading of a joint, in and of itself, causes fracture of bone rather than cartilage
[7]. In contrast, subfracture impulsive loads cause microinjury of both the
subchondral bone and articular cartilage that may exceed the ability of the
joint to repair the damage.
The capacity for intrinsic repair of damaged articular cartilage is limited,
but cells that are extrinsic to the cartilage can provide a mechanism for re-
pair if the local environment permits [5]. Although the new cartilage they
produce, fibrocartilage, is not histologically, biochemically, or biomechani-
cally comparable to normal hyaline articular cartilage, it nonetheless per-
mits normal joint function in the presence of physiologic loading, prevents
further deterioration, and, most important, permits the patient to function
asymptomatically [5]. Data indicate that joint healing in OA depends both
on a source of cells and on the normalization of intra-articular stress and
movement of the joint [8]. Many investigators do not recognize sufficiently
the large role that reduction of the excessive levels of mechanical stress plays
in promoting the healing response in OA or that the most important out-
comes of the healing process are the relief of joint pain and improvement
of function, not the histologic appearance of the articular surface.
The importance of abnormal joint mechanics in the etiopathogenesis of OA
cannot be overemphasized. For example, in the classical experimental canine
model of knee OA that is induced by transaction of the anterior cruciate liga-
ment, OA will not develop if loading of the unstable limb is restricted markedly
by immobilization immediately after instability is created [9]. Immobilization
effectively reduces the instability, the loss of the load-bearing surface, and the
stresses on joint tissues that exceed physiologic limits.
On the other hand, when the stresses on joint tissues exceed physiologic
limits (because of obesity, varus-valgus malalignment, meniscus damage,
developmental or genetic factors, or other causes) and overwhelm the pro-
tective mechanisms, OA ensues. In such cases of OA, even though the loads
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 535

on the joint may be physiologic, the accumulation of abnormal biomaterials

in joint tissues, such as polymers of homogenistic acid in ochronosis, results
in abnormal tissue mechanics and, eventually, in breakdown of the joint.
The cells in cartilage and bone in an OA joint are essentially normal, but
in highly loaded areas of the joint they seem abnormal because they literally
are being crushed. If the abnormal joint mechanics can be restored to a phys-
iologic range, however, joints can heal [5]. Although the process is slow, the
fibrocartilage and woven bone that have been mistakenly characterized as
a failed attempt at healing are clinically functional and may be associated
with symptomatic relief. If normal joint mechanics can be sustained, these
transitional tissues eventually may remodel into normal hyaline cartilage
and trabecular bone [5].
Eventually, all the tissues of the joint are involved in OA, including syno-
vium, periarticular muscle, nerves, ligaments, and, if present, meniscus,
as well as the articular cartilage and subchondral bone. It follows that the
emphasis of clinical and research in OA should be on preventing and treat-
ing the increased intra-articular stress that is causing joint damage rather
than, as now, almost exclusively on the loss of articular cartilage.
Because OA is more than a cartilage problem [10], it is notable that
Heberden’s nodes (which generally are interpreted as a manifestation of OA
of the distal interphalangeal joints and have been reported to be a risk factor
for incident knee OA [11] and for the progression of established knee OA
[12]) can be associated with a variety of pathologies of the distal interpha-
langeal joint [13]. A slide in the 1972 Arthritis Foundation Clinical Teaching
Collection depicting a histologic section through a Heberden’s node from
a patient who has nodal OA illustrates this point (Fig. 1) [14].

Fig. 1. Histologic section through a distal interphalangeal joint of a patient who has nodal
osteoarthritis. A prominent dorsal osteophyte is present, but there is no thinning of the articular
cartilage or loss of surface integrity and no thickening of the subchondral bone, which are
pathognomonic features of osteoarthritis. (From The American College of Rheumatology
clinical slide collection. Atlanta (GA): American College of Rheumatology; Ó 1972–2004
American College of Rheumatology Slide Collection. Used wih permission.)
536 BRANDT et al

A prominent osteophyte that is present at the articular margin of the distal

phalanx is unaccompanied by changes in the surface integrity or in the thick-
ness of the load-bearing articular cartilage or by thickening of the subchon-
dral bone, changes that, in combination, are pathognomonic for OA. The
slide is consistent with observations that osteophytes are not pathognomonic
for OA. Hernborg and Nilsson [15], for example, concluded in 1973 that, in
the absence of other bony changes (ie, subchondral sclerosis or cysts), knee
osteophytes might be a manifestation of aging rather than of OA.
The authors suggest that the long-standing inability to manage OA pa-
tients effectively with pharmacologic approaches is the result of a persistent
overemphasis on articular cartilage as the major focus of OA damage and
on cartilage in the progression of joint failure. Excessive emphasis also
has been placed on the inflammatory phase of OA. It has been insufficiently
appreciated that the synovial joint is an organ, that OA is organ failure, and
that this organ failure can be initiated by abnormalities arising in any of the
organ’s tissues. The authors believe the failure to focus on the reduction in
the contact area in the joint and on aberrant joint loading has misdirected
the therapeutic efforts in OA.
The authors also believe that effecting cures that aim at the causes of OA
will require the understanding that OA can be the result of traumatic joint
damage or inherited predilections. Many of the cases of OA that do not
follow significant (macro) joint injury may be related to abnormalities in
genes that create or promote the development of structurally abnormal
joints. These abnormal joints result in aberrant loads and/or increased
intra-articular stress, leading to cumulative tissue microdamage and remod-
eling that are detrimental to joint function [16].
Not all abnormal joints develop OA, however. Among subjects who had
hip dysplasia, slipped capital femoral epiphysis, or Legg-Calve-Perthes dis-
ease in childhood, the frequency of OA at a 30-year follow-up evaluation
was 60% to 70%, not 100% [17]. Although all subjects had hips that
were at increased risk for developing OA, the likelihood that a predisposed
individual would develop OA presumably was affected by the severity of the
structural abnormality, the amount of repetitive unprotected loading, and
the adequacy of the mechanisms that normally protect joints from excessive
mechanical stress.

Articular cartilage healing and remodeling: mechanobiology

of the chondrocyte
The remodeling of connective tissues requires removal of damaged matrix
to allow for its replacement. Inflammation and alterations of cell metabo-
lism are an essential part of this healing process. In a normal joint, chondro-
cytes in the articular cartilage are subjected to physiologic dynamic and
static compressive and deep shear stresses. When explants of normal
articular cartilage were subjected to noninjurious loading (ie, loading that
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 537

did not result in loss of cartilage matrix molecules into the culture medium
or damage to the collagen network), the physical perturbation of the chon-
drocytes was associated with changes in expression of genes for aggrecan,
collagen, matrix metalloproteinases, growth factors, and cytokines that
were transduced into metabolic responses [18].
The rate of loading may be more injurious to cartilage than the magnitude
of the load. When normal rabbits were subjected to repeated acute (50 milli-
seconds, onset to peak) impulsive loading of the knee (ie, square-wave load-
ing), the articular cartilage and subchondral bone were damaged. When
applied more gradually, loads of greater magnitude (500 milliseconds, onset
to peak; ie, sine-wave loading) had no adverse effect [19]. Rapid delivery of
load does not permit sufficient time for periarticular muscle, the major shock
absorber protecting the joint, to prepare for and absorb the load [20].
In vitro, a single injurious compression of a cartilage explant resulted in
a 250-fold increase in the expression of the gene for matrix metalloprotei-
nase-3 (stromelysin), a 40-fold increase in the expression of the gene for
ADAMTS-5 (aggrecanase), and a 12-fold increase in the expression of the
gene for tissue inhibitor of matrix metalloproteinases [21]. Changes of this pro-
portion in the quantities of the relevant proteins that are expressed could lead to
breakdown of the articular cartilage matrix. The authors consider the response
of the cartilage in this experiment to be evidence of an attempt by the chondro-
cytes to repair obvious damage induced in vitro, which was manifest by the loss
of cartilage matrix molecules into the culture medium and damage to the col-
lagen network. They suspect the metabolic responses of cartilage to noninjuri-
ous loading represent a similar phenomenon but that the criteria for identifying
cartilage injury were not sufficiently sensitive to detect microdamage.
The effect on joint cartilage of loading forces and of cytokines (eg, tumor
necrosis factor, interleukin-1) may be additive. When explants of normal ar-
ticular cartilage that had been damaged mechanically were co-cultured with
normal joint capsule/synovium, they released higher levels of aggrecanase-
generated products of aggrecan and exhibited higher levels of aggrecanase
than seen in normal cartilage or in damaged cartilage that was cultured in
the absence of joint capsule/synovium [22]. These observations suggest
that, beyond the direct effects of the mechanical insult to the cartilage itself,
capsule/ synovium of the OA joint may contribute to the degradation of the
damaged cartilage in an attempt to heal the joint.
Although dose–response curves that are relevant to loading of joint car-
tilage in vivo are not well defined, the data suggest that the breakdown of
OA cartilage is mediated by matrix metalloproteinases whose production
is stimulated by cytokines, such as interleukin-1 and tumor necrosis factor,
and that these cytokines could be produced by the chondrocytes themselves
in response to mechanical loads. It is reasonable to consider that the process
may be driven by abnormal mechanical stresses on the joint (eg, arising as
a result of muscle weakness or neuropathy) that interfere with protective
muscular reflexes [23] or by ligamentous instability.
538 BRANDT et al

Changes seen in organ culture in vitro are not identical to what happens in
an OA joint in vivo, however. As indicated earlier [7], excessive loading of
a joint will lead to fracture of the bone before it produces obvious damage
to the articular cartilage. Although metalloproteinases can weaken the
mechanical integrity of articular cartilage in OA, mechanical factors seem
to play the major role; pharmacologic inhibition of metalloproteinases has
not been uniformly successful in halting the progression of OA [24,25]. Based
on the readily apparent shards of articular cartilage embedded in OA syno-
vium [26] and the duplication and advance of the tidemark [27,28], much
of the loss of articular cartilage in OA seems to be caused by the breaking
off of enzymatically weakened segments of the joint surface, like icebergs
from a glacier. Although vertical splits (fibrillation) in the proteoglycan-
depleted cartilage can persist without fragmentation of the joint surface,
the thinned cartilage also is subject to deep horizontal splits (Fig. 2) [29].
When these deep horizontal splits join with the common vertical splits, the
result is the breaking off of cartilage shards that may be detected in the syno-
vial fluid before they become incorporated in the synovial membrane, where
they incite inflammation. In addition, endochondral ossification, caused by
the reactivation of the secondary center of ossification, moves toward the
joint surface, duplicating and advancing the tidemark and gradually thinning

Fig. 2. Histologic section of articular cartilage from a rabbit whose hind limb has been
subjected to repetitive impulsive loading (see [19]). In addition to vertical splits (fibrillation)
at the surface, horizontal tears are apparent running near the base of the cartilage. In osteoar-
thritis, when such tears join with the vertical fissures, they result in fragmentation of the joint
surface and are responsible for the presence of particulate cartilage debris in the synovial space
that may become incorporated into the synovial membrane and cause synovitis (as may soluble
breakdown products of cartilage matrix molecules). Safranin O-fast green. Original magnifica-
tion  25. (Courtesy of D.B. Burr, PhD, Indianapolis, IN.)
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 539

the cartilage from below [27]. Most of the loss of articular cartilage in OA
seems to be attributable to these two processes.

Why does any of this matter? Lack of relevance of chondroprotective drugs

to the underlying process
Understanding the etiopathogenesis of OA matters because of the impli-
cations for treatment of patients. Because the synovial inflammation in OA
is secondary to mechanical damage to the cartilage and bone, it stands to
reason that nonsteroidal anti-inflammatory drugs (NSAIDs), although
they may be effective symptomatically, cannot arrest the process. Efforts
to develop disease-modifying OA drugs (which formerly were called ‘‘chon-
droprotective drugs’’ and now also are called ‘‘structure-modifying OA
drugs’’) that can interrupt or reverse the pathogenetic processes underlying
OA have not, as noted earlier, been unambiguously effective [24,25]. The au-
thors believe the reason for this lack of success is the failure to appreciate
sufficiently that OA, from its earliest stages and in whichever joint it occurs,
is mechanically induced [10,25].
A variety of mechanical abnormalities can trigger the many processes
involved in joint repair and attempts to contain the mechanical insult. With-
out a return to mechanical normality, however, attempts at healing will fail.
Drugs cannot accomplish this return to mechanical normality. As long as
the joint remains in the same adverse mechanical environment that created
the trouble in the first place, it seems highly unlikely that a drug that inhibits
a specific enzyme or cytokine in the pathways of cartilage breakdown or that
further stimulates the already increased synthesis of cartilage matrix mole-
cules by the chondrocytes will solve the problem of OA. Indeed, if the
increased joint stresses are normalized, such drugs may be superfluous. In
addition, as explained later, because the subchondral bone is critical for
containment of the mechanical abnormalities that damage the articular
cartilage, emphasis on cartilage repair alone is likely to be futile.
The authors believe that the genes whose identification will be required to
allow investigators to get at the root problems in OA are not the genes that
regulate chondrocyte metabolism. Rather, they will prove to be genes that
control the developmental and congenital deformities that lead to a reduction
in the habitually loaded area of the joint surface and result in the concentra-
tion of peak dynamic loads or the genes that create micro-incoordination
during physical activity [6].
The authors do not think that chondroprotective drugs are the answer to
OA. Instead, just as an increase in intra-articular stress can cause OA, low-
ering intra-articular stress can result in healing of OA [5]. Notably, healing
can be brought about only by mechanical means. Drugs have not been
shown to accomplish this healing, and arthroplasty (essentially, an amputa-
tion of the joint and replacement with a prosthesis) is the popular current
treatment for OA. With a carefully executed osteotomy, however, the
540 BRANDT et al

load-bearing surface may be increased (Fig. 3) and/or the force of a selected

muscle acting across a joint can be reduced by increasing the leverage or by
weakening the muscle by lengthening it [30,31].

Why doesn’t everyone develop progressive osteoarthritis? Mechanisms

protecting the joint from microdamage
In normal walking, three to four times the weight of the body is transmit-
ted through the knee; during a deep knee bend, the patellofemoral joint is
subjected to a load 9 to 10 times body weight. Adaptive mechanisms must
protect the joint from these physiologic loads. Although the bulk properties
of articular cartilage would make it an excellent shock absorber, at most
sites it is too thin to serve as an adequate shock-absorbing structure in
a joint. The periarticular muscles and subchondral bone provide the addi-
tional protective mechanisms that are needed.

Periarticular muscle as a shock absorber for the joint

The most active shock-absorbing mechanisms for joints involve the use of
muscles and joint motion. Although muscle contraction can move a joint,
muscles also can act like large rubber bands. When a slightly stretched mus-
cle is subjected to greater stretch by the movement of a joint, it can absorb

Fig. 3. Radiographs of a patient who has hip dysplasia that led to osteoarthritis. The radio-
graph on the left was obtained preoperatively. The film on the right, of the same hip, was
obtained 7 years after a successful osteotomy. At surgery, the alignment of the femur was
altered so that the large, inferomedial beak osteophyte was shifted to the load-bearing region
of the joint. Note the marked decrease in subchondral sclerosis and the widening of the joint
space after surgery. (From Maquet P. Biomechanics of the hip as applied to osteoarthritis
and related conditions. Berlin: Springer; 1985. p. 114. With the kind permission of Springer
Science and Business Media, and of Dr. Maquet.)
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 541

a large amount of energy. Most of the muscle activity generated during

activity is not used to propel the body forward but to absorb energy to
decelerate the body [32].
For example, when a person jumps off a ledge or table, he or she normally
lands on the toes, comes down on the heels, and then straightens the flexed
knees and hips. During this smooth action the muscles perform negative
work; that is, they absorb energy. As one dorsiflexes the ankles, one stretches
the calf muscles; as one straightens the knees and hips, one stretches the ham-
strings. The amount of energy absorbed as a result of this stretching is enor-
mous [33]. The energy produced by normal walking is sufficient to tear all the
ligaments of the knee; that this tearing does not occur routinely attests to the
importance and effectiveness of the active energy absorption by the muscles
that surround the joints and cushion them from mechanical stress.
Small, unexpected loads for which one is unprepared are much more
damaging to joints than large ones that have been anticipated. Consider
what happens when one comes down stairs, misjudges a step, and abruptly
skips a step. Because the muscles are not prepared to accommodate the load,
one feels a sharp jolt. To prepare a neuromuscular reflex to handle an
impact load requires approximately 75 milliseconds. Therefore, an unex-
pected fall of very brief duration (eg, of only about 1 inch) does not afford
sufficient time to bring protective muscular reflexes into play [20]. Under
such conditions, the load is transmitted to the cartilage and bone. In con-
trast, during a fall from a slightly greater height, sufficient time is available
to activate the appropriate reflexes, so that the lengthening of the muscles
that cross the joint protects the joint’s articular cartilage by absorbing the
energy of impact [33].

The quadriceps muscle: implications for knee osteoarthritis

The quadriceps muscles serve as a good example of the processes
described in the previous section. Muscle weakness (which may occur in
association with knee OA) and an increase in the latent period of the reflex,
which may occur with peripheral neuropathy resulting from aging or other
causes, can reduce the effectiveness of this shock-absorbing mechanism.
The quadriceps muscle is the major antigravity muscle of the lower
extremity. It protects the knee from mechanical damage by serving as
a brake on the pendular action of the lower limb during ambulation, thereby
minimizing the forces generated with heelstrike. In addition, the quadriceps
muscle is important in stabilizing the knee. Hence, quadriceps weakness may
generate abnormal mechanical stresses on the joint.
After a femoral nerve block (to paralyze the quadriceps temporarily), the
load rate in normal subjects who had no force-transient profile during gait
increased more than twofold to approximately 150 times body weight/second
[34]. The data suggest that a force transient can be caused by failure to decel-
erate the lower extremity before heelstrike. In normal individuals, minor
542 BRANDT et al

incoordination in muscle recruitment, resulting in failure to decelerate the leg

before heelstrike, may generate rapidly applied impulsive forces at the knee
that are as high as 65 times body weight/second [6]. As noted earlier, in rab-
bits repetitive rapidly applied loading of the knee for which the animals could
not prepare resulted in the breakdown of articular cartilage and bone,
whereas loads of greater magnitude, applied more gradually, were innocuous.
Whether the micro-incoordination of muscular control in some normal
humans that results in the generation of a heelstrike transient at the knee
with walking is a risk factor for knee pain or for structural changes of knee
OA remains to be established by prospective longitudinal studies.
Although the periarticular muscles serve a primary motor function, Hurley
[35] has emphasized the importance of the sensory function of muscle and of
the proprioceptive impulses that originate in muscle and are transmitted to
the central nervous system. Data suggest that muscle weakness caused either
by reflex inhibition of muscle contraction or by intra-articular pathology may
result in joint degeneration.
Quadriceps weakness is common in patients who have knee OA. In these
patients quadriceps weakness has generally been thought to arise as a conse-
quence of the pain that occurs with loading of the arthritic joint; it was pos-
tulated that the pain caused the patient to minimize load bearing, thereby
leading to atrophy. Quadriceps weakness also may exist also in subjects
who have knee OA but who have no history of joint pain and in whom
quadriceps muscle mass is not diminished but is normal or even increased
(as a result of obesity) [36]. Longitudinal studies suggest that, in addition
to arising from painful knee OA, quadriceps weakness itself may be a risk
factor for structural damage [36–38]. Baseline knee extensor strength was
significantly lower among women who had no radiographic evidence of
knee OA at the initial examination but who had developed OA changes
some 30 months later than in women who did not develop radiographic
changes of OA [36]. When the presence of knee OA (based on radiographic
changes, with or without knee pain) as a function of sex, body weight, age,
and lower extremity strength was modeled, it was found that each 10-lb/ft
increase in knee extensor strength was associated with a 20% reduction in
the odds of developing radiographic OA and with a 29% reduction in the
odds of developing symptomatic knee OA. A relatively small increase in
strength (approximately 20% of the mean for men and 25% for women)
was predicted to result in a 20% to 30% decrease in the odds of having
knee OA [36].
Although the question has not been examined, the shock-absorbing func-
tion of periarticular muscle in protecting joints other than the knee probably
is important also. The expression of OA (eg, with respect to the frequency
and severity of joint pain and the rate of progression of joint damage)
may vary at different joint sites, and some of this variability may be influ-
enced by the adequacy of the physiologic shock-absorbing mechanisms
that protect the joint.
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 543

Subchondral bone as a passive shock absorber for the joint

The normal joint
Normally, in the unloaded state, the opposing surfaces of a joint are
incongruent. Under load, the cartilage and bone on both sides of the joint
space deform, maximizing the contact area and thereby minimizing the
stress within the cartilage [39]. This normal deformation of the articular car-
tilage when subjected to a compressive load provides the self-pressurized
hydrostatic weeping lubrication needed for effortless motion. With increas-
ing load, however, cartilage deformation alone is insufficient; deformation
of the underlying bone must occur also and, under high loads, is more
important than deformation of cartilage in reducing stress.
As noted previously, joint damage caused by excessive loading is related
to the rate, and not only to the magnitude, of loading, because of the limi-
tation of the chondroprotective effect of the shock-absorbing capacity of
subchondral bone [40]. In brief, normal subchondral bone is viscoelastic;
that is, it deforms less, or becomes stiffer, when load is applied rapidly
than when it is loaded more gradually. Rapid loading of the joint does
not allow time for the viscoelastic flow of interstitial fluid out of the bone,
which would absorb the energy transmitted and protect the cartilage matrix
and chondrocytes. Therefore, if deformation under an impulsive load is
restricted, the cartilage cannot conform completely to the load, the size of
the contact area is restricted, and high stresses are generated in the cartilage
matrix. This process is exacerbated by thinning of the cartilage and is the
reason that rapid impulsive loading eventually is detrimental to the integrity
of the articular cartilage.
The cancellous subchondral bone, although 10 times stiffer than the carti-
lage, is much softer than cortical bone and serves as a major shock absorber.
Because the cartilage is too thin to function effectively as a shock absorber, the
subchondral bone, by providing a pliable bed that absorbs energy, protects it
from damage caused by repetitive impulse loads [39], which are capable of
fracturing the subchondral trabeculae.
Theoretically, the transfer of load from the articular surface to the diaph-
yseal cortex should create large shear stresses in the subchondral bone,
particularly under the edges of the contact area [41]. Because of the undula-
tions of the tidemark and the osteochondral junction, the interposed zone of
calcified cartilage, whose stiffness is intermediate between that of the over-
lying hyaline cartilage and that of the underlying subchondral bone [42],
tends to transform these shear stresses into compressive and tensile stresses
[43] that the articular cartilage is better able to withstand.
Joint lubrication is so effective that shear forces along the surface of the
cartilage are unlikely [44]. Furthermore, by constraining the radial deforma-
tion of the cartilage under load [45], the subchondral bone raises the thresh-
old for cartilage damage. For example, grasping with the hand or, in most
normal individuals, walking does not generate impulsive loads on contact of
544 BRANDT et al

the joint surfaces [6]. Articular cartilage can withstand up to five times the
peak deformation that occurs with walking, suggesting that normal
subchondral bone can protect it from all but impulsive loads. Bone, like car-
tilage, is viscoelastic; that is, the fluid in the tissue acts to dampen the effects
of loading.
With rapidly applied impulsive loads, however, the viscoelasticity of the
tissue becomes problematic. Viscoelastic damping requires time to have an
effect: fluid must flow. About one third of normal adults are afflicted with
micro-incoordination [6], and in these persons the important muscle-based
protective mechanisms needed to dampen the forces of joint loading are
not fully coordinated. These individuals therefore subject their knees to
impulsive loading during walking. The consequence of impulsive loading
is that the subchondral bone, which is only one tenth as stiff as hyaline
cartilage, takes the brunt of the blow and, along with the calcified cartilage,
sustains microdamage. If this damage is repetitive, it will cause reactivation
of the secondary center of ossification and remodeling of the subchondral
bone, with an advance of the tidemark leading to thinning of the articular
cartilage [27].

The osteoarthritic joint

Normal subchondral bone attenuates loads through the joint more than ei-
ther the articular cartilage or surrounding soft tissues [46]. In a normal joint,
subchondral bone absorbs up to 50% of the load, and the cartilage absorbs
only 1% to 3% [46,47]. In an OA joint, however, the sclerotic subchondral
bone is less able to absorb and dissipate the energy of an impulsive load,
increasing the force transmitted through the joint. Therefore, the OA knee
absorbs only about half as much load as a normal knee [47]. As Burr [48]
noted, the total volume of subchondral trabecular bone in OA increases by
an average of 10% to 15%, chiefly because of the thickening of the trabeculae
and some increase in the number of trabeculae and reduction in trabecular sep-
aration. This process is reflected by the subchondral sclerosis that is a cardinal
radiographic feature of OA and by an increase in the apparent density (bone
volume/total volume of the tissue) of OA bone in comparison with normal
bone. Because the subchondral bone in OA is remodeling actively in response
to the increased mechanical stress, however, much of the newly formed bone
does not have sufficient time to mineralize fully. Therefore, from a material
standpoint, this bone is less highly mineralized and is less stiff than bone
from age-matched nonarthritic controls [49,50]. To retain a normal degree
of tissue stiffness, or even that seen in osteoporosis, the volume of subchondral
bone in an OA joint must increase markedly, to at least twice the thickness in
a normal joint [48].
An earlier hypothesis proposed that stiffened subchondral bone, present
in the OA joint as a consequence of the healing of cumulative microfrac-
tures, drives the destruction of the overlying articular cartilage. This sugges-
tion was not supported by finite element models, which have shown that
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 545

even a marked increase in the density of subchondral bone would increase

the mechanical stress in the overlying cartilage only modestly [51]. The
amount of stiffening of the bone required to increase stresses in the cartilage
significantly is far beyond that which is likely to occur in vivo. Although
stiffening of subchondral bone may be an important feature of OA, these
data suggest that, on a mechanical basis, stiffening alone is not sufficient
to account for the destruction of the articular cartilage [28]. It is the thinning
of the cartilage as the subchondral bone thickens and advances toward the
joint space that eventually causes the cartilage to fragment.
Microcracks in the subchondral bone or calcified cartilage stimulate focal
remodeling and account for the increased vascularity in OA joints. Micro-
cracks are found routinely in calcified cartilage from femoral heads of
middle-aged nonarthritic humans and are associated with foci of remodeling
in OA cartilage [48]. Single or repetitive high-impact loads have been shown
to cause microcracks [52], which are followed by remodeling of the subchon-
dral bone and degeneration of the overlying cartilage. Thus, microdamage
in the calcified cartilage caused by mechanical stress, and the ensuing endo-
chondral ossification, may play a vital role in the pathogenesis of OA.
With remodeling and disruption of the articular plate by microfractures,
bone necrosis, as reflected by empty lacunae devoid of osteocytes, is seen in
association with extensive resorption, repair, and apposition of new bone on
the existing subchondral trabeculae. Scintigraphy may show a generalized
subarticular zone of increased uptake, whose presence correlates with pain
on loading of the joint, subchondral sclerosis, and progression of the struc-
tural damage of OA [53]. The appearance, however, differs from that of pri-
mary osteonecrosis, in which a sequestrum of subchondral bone underlies
viable articular cartilage. The contribution of this secondary osteonecrosis
to the collapse of the articular surface in OA is unclear, but it has been con-
sidered to be related to the occlusion of small intramedullary arteries and
has been reported to occur in nearly one of every six femoral heads resected
surgically for advanced OA [54].
Whether changes in the subchondral bone precede or follow those in the
overlying cartilage in OA is unclear. It is likely that both sequences occur.
Even when bony changes are not involved in the initiation of cartilage dam-
age, these changes are critically important in the progression of cartilage
breakdown in OA. Evidence that changes in bone may precede those in
the cartilage is provided by the results of a study in a rabbit model of
OA, in which subfracture loads were applied to the patellofemoral joint
[55]. Significant reduction in cartilage stiffness was not seen until 12 months
after the injury, whereas a progressive increase in the thickness of the sub-
chondral plate was apparent by 6 months. In contrast, in other animal
models of OA, thickening of the subchondral plate was not an early event
but was associated with subsequent progression of cartilage damage
[56,57]. In dogs in which knee OA was induced by transection of the anterior
cruciate ligament, treatment with an antiresorptive drug (bisphosphonate)
546 BRANDT et al

sharply reduced turnover of the subchondral bone, but it did not affect bio-
chemical, metabolic, or morphologic changes in the cartilage [58]. Further-
more, although the drug inhibited the formation of subchondral bone,
which is coupled to bone resorption, it had no effect on osteophytes.

Etiopathogenesis of osteoarthritis pain

For the patient and clinician, the essential problem associated with OA is
joint pain. Indeed, if OA were not painful, it would receive little attention.
Liang [59] succinctly captured the issue: ‘‘[X]-rays don’t weep.’’ Patients weep.
The following discussion places that statement into context. Among
people who have OA, those who have more severe radiographic changes
are somewhat more likely to have difficulty with mobility, but the radio-
graphic progression of OA usually is slow, and the course is unpredictable
[60]. Symptoms wax and wane independently of radiographic progression
[61]. Among people who have far advanced radiographic changes of hip
OA, nearly half may not have hip pain [62,63]. On the other hand, the
overall prevalence of radiographic OA is very high: the Framingham Study
suggested it occurs in one third of all people age 63 years and older [64].
Even if joint pain occurs in only a minority of these individuals, the absolute
numbers are so large that they present a major public health problem with
huge medicoeconomic and socioeconomic costs.
Given the prevalence of asymptomatic radiographic OA, it is notable that
criteria for case definition traditionally have relied on the presence of radio-
graphic features of OA. The use of radiographic criteria to define cases of
OA for clinical studies has serious limitations, however. The association
between radiographic changes of OA and reported pain in the hip or knee
may be statistically significant in groups of people but, as indicated earlier,
in an individual patient the correlation between the severity of radiographic
changes (ie, structural damage) and the severity of symptoms often is poor.
At the community level, many people who have severe radiographic OA
have no symptoms.
The real problem is not OA but rather is painful OA, and current reports
of OA research often contain a statement to the effect that OA is a disease not
merely of cartilage but of the entire joint. The large amounts of time, money,
and brainpower that are invested in efforts to develop a chondroprotective
drug, to image miniscule (and clinically meaningless) changes in articular
cartilage, and to identify markers of cartilage damage in body fluids suggest
that this proposition is not really accepted. Current textbooks discuss at
length the biochemical changes that underlie the breakdown of articular
cartilage in OA [65–67], but biochemists and molecular biologists studying
cartilage or bone from OA joints, investigators imaging OA joints, and those
assaying biologic fluids for molecules derived from OA joints have yet to
explain why some OA joints are painful and others are not. In patients
who have knee OA, anxiety, depression, and quadriceps weakness may be
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 547

better predictors of pain than the severity of radiographic changes [68].

Understanding the etiopathogenesis of OA pain may be much more relevant
than understanding the mechanisms of articular cartilage breakdown in this
disease.
The pathogenesis and the tissue origins of OA pain may vary from
patient to patient and in the same patient from visit to visit. A variety of
evidence points to the synovium and subchondral bone as the major sources
of joint pain in patients who have OA, but ligaments, menisci, periarticular
muscles, entheses, and the joint capsule all can be painful.

Synovitis
The synovial membrane from patients who have advanced OA commonly
exhibits hyperplasia of the lining cell layer and focal infiltration of lympho-
cytes and monocytes. In advanced OA the intensity of the synovitis may
resemble that in rheumatoid arthritis. Synovitis in OA may be caused by
phagocytosis of wear particles of cartilage and bone from the abraded joint
surface [26,69,70]; by release from the cartilage of soluble matrix macromol-
ecules [71] (eg, proteoglycans, collagen, fibronectin fragments); or by the
presence of crystals of calcium pyrophosphate dihydrate or calcium hydroxy-
apatite [72]. In some cases, immune complexes containing antigens derived
from the cartilage matrix may be sequestered in collagenous tissue of the
joint, such as meniscus, leading to chronic low-grade inflammation [73].
Earlier in the course of OA, however, the synovium, even from symptom-
atic patients who have full-thickness ulceration of the articular cartilage,
may be histologically normal, suggesting that the early pain in those cases
is not caused by synovitis [74]. Conversely, the severity of articular cartilage
damage and of synovitis may be as great in patients who have knee OA but
no joint pain as in those who do have knee pain.
In cross-sectional MRI analyses of subjects who had knee OA, synovial
thickening was much more common in subjects who had pain than in those
who were asymptomatic and, among those who had knee pain, was associ-
ated with more severe pain [75]. Furthermore, in a 30-month longitudinal
study of patients who had symptomatic knee OA [76], changes in synovitis,
as graded by MRI, correlated only modestly with changes in knee pain. The
relatively weak correlation suggests that synovitis was not the only or even
the major cause of the joint pain. Furthermore, pain was not correlated with
the loss of articular cartilage in either the tibiofemoral or patellofemoral
compartment, and changes in synovial effusion were not correlated with
changes in pain. In contrast, in a sample of symptomatic subjects from
the Osteoarthritis Initiative, Lo and colleagues [77] found that maximal joint
effusion scores on MRI were highly associated with knee pain even after
adjustment for bone marrow lesion (BML) scores, suggesting that effusion
(a manifestation of underlying synovitis) was associated independently
with knee pain.
548 BRANDT et al

Among the reservations the authors expressed earlier about the success of
pharmacologic modification of joint damage in OA without the establish-
ment of a more favorable local mechanical environment is the uncertainty
that a chondroprotective drug would ameliorate symptoms. If such an agent
stabilized the cartilage surface, it might reduce the component of joint pain
caused by synovitis resulting from the breakdown of cartilage and bone.
There is no reason, however, to believe the agent would reduce pain that
is caused by the reaction to phlogistic breakdown products that already
had been incorporated by the synovium. In addition, in the face of increased
intra-articular stress, it is unlikely that pharmacologic chondroprotection is
feasible.

Subchondral bone
In the early 1970s, Arnoldi and colleagues [78,79] emphasized the impor-
tance of increased intraosseous pressure as a cause of pain in OA. These
investigators demonstrated marked increases in intramedullary pressure
and stasis of medullary blood flow in patients who had hip OA, presumably
caused by the distortion of blood flow through subchondral trabeculae that
had been thickened as a result of remodeling in response to abnormal stress.
Furthermore, normalization of the hemodynamic changes by an osteotomy
(ie, an operation that, without entering the joint, cuts the bone and fixes it
into an alignment that relieves mechanical stress across the joint by decreas-
ing the resultant force on the joint and/or increasing the surface area avail-
able to carry that force) or by any procedure that transects the metaphyseal
bone can alleviate joint pain promptly.
In 1980, Arnoldi and colleagues [80] described 25 patients who had pain in
the hip or knee at rest caused by OA or by intraosseous engorgement-pain syn-
drome who exhibited venous stasis on intraosseous phlebography, increased
intraosseous pressure in the intramedullary space near the painful joint,
and high uptake of 99mtechnetium polyphosphate on bone scintigraphy. In
patients who had other types of pain, these correlations did not exist. The
authors suggested that the similarity of the three findings in patients who
had intraosseous engorgement-pain syndrome and those who had OA
reflected a common pathogenetic mechanism for the patients’ pain.
Notably, McAlindon and colleagues [81], in a 1991 study that correlated
findings on MRI, scintigraphy, and radiography in 12 patients who had
knee OA, found that nine OA knees exhibited diffuse loss of the medial or
lateral tibiofemoral subchondral marrow signal on the proton-density image,
with corresponding hyperintensity on the short tau inversion recovery
sequence, and reported that this MRI abnormality was associated very
strongly with an extended pattern of isotope on bone scintigraphy.
In a cross-sectional scintigraphic study of 100 patients who had knee OA
recruited from a rheumatology clinic, McCrae and colleagues [53] found
a very strong correlation between a pattern of generalized uptake of the
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 549

isotope around the joint on the delayed (bone-phase) scan and knee pain.
Similarly, a very strong correlation was observed between isotope retention
in the subchondral bone in late-phase scans and subchondral sclerosis on the
radiograph. Furthermore, Dieppe and colleagues [82] documented the pre-
dictive value of the bone scan for radiographic progression of knee OA, as
reflected by joint-space narrowing. In scans that did not show focal areas of
isotope retention, no progression was noted over a 5-year follow-up period.
In contrast, progression of joint-space narrowing was seen in approximately
50% of patients who had OA and who had a ‘‘hot’’ knee scan at baseline.
Since these reports by Arnoldi and colleagues [80], however, osteotomy
largely has been replaced by arthroplasty as a surgical treatment for symp-
tomatic OA. Nonsurgical treatment of OA pain continues to be based
chiefly on systemic pharmacologic therapy with analgesics and nonselective
or selective NSAIDs that fail to take into account the altered vascular
physiology of the joint.
A recent preliminary report by Hunter and colleagues [83] confirms a link
between elevated intraosseous pressures and BMLs, which are seen com-
monly on MRI of OA knees as focal areas of increased signal in the subchon-
dral marrow in fat-suppressed T2-weighted images. The lesions that now are
called ‘‘BMLs’’ are consistent with the marrow lesions that McAlindon and
colleagues [81] described in 1991 and showed to be associated strongly with
isotope retention on bone scintigraphy. Histologic examination of these
lesions, which erroneously have been called ‘‘bone marrow edema,’’ has
shown them to be foci of fibrosis and of osteonecrosis and bone remodeling
[84]. BMLs are not specific for OA and may be seen with insufficiency frac-
tures, osteonecrosis, and a variety of other conditions. In patients who
have knee OA, they have been associated with varus-valgus malalignment
[85] and the progression of structural damage [81,85], and, in some studies,
with joint pain [53,86].
Using a dynamic contrast-enhanced MRI sequence, Hunter and col-
leagues [83] found that BMLs in patients who had advanced knee OA
were sites of venous hypertension that showed reduced runoff of contrast
material in comparison with the surrounding tissue. Results were consistent
with reduced perfusion, intraosseous venous hypertension, and increased
permeability at sites of BMLs. It is not known whether pharmacologic
agents acting directly on vascular flow would be efficacious in treatment
of OA pain.
However, in a recent uncontrolled trial in 104 patients who had painful
BMLs from a variety of causes, including OA, Meizer and colleagues [87]
found that parenteral administration of iloprost, a stable analogue of pros-
tacyclin, significantly improved pain and decreased the size of the BMLs.
Aigner and colleagues [88] compared results over 11 months in patients
who had bone marrow edema syndrome (with BMLs on MRI but not
necessarily with OA) who were treated with iloprost and in a control group
that was treated with core decompression of the femoral head. Clinical
550 BRANDT et al

improvement and improvement in the MRI changes were marked in both

groups and were as good, or better, with iloprost as with core decompres-
sion. It would be of interest to know whether iloprost or other vasoactive
agents are of symptomatic benefit in patients who have OA and, if so,
whether improvement can be predicted by the presence of a BML on
MRI. It also would be important to know how long such changes persist.
In a cross-sectional study, Lo and colleagues [77] found that the maximal
BML score was associated highly with knee pain even after adjustment for
the synovial effusion score, suggesting that each is independently associated
with joint pain. In a 15-month longitudinal study of the relationship between
BML and knee pain in older patients who had knee OA or were at high risk
for developing knee OA, Felson and colleagues [89] reported that an increase
in the BML volume score at 15 months, relative to the baseline score, was
much more common among those who had no knee pain at baseline but
who developed knee pain during the study than in those who did not develop
knee pain.
In a preliminary report of a cross-sectional study of subjects who had
symptomatic OA and were enrolled in the Osteoarthritis Initiative, a multi-
center study of the natural history of knee OA, Lo and colleagues [90] found
a strong association between maceration of the meniscus and BMLs. Large
BMLs in the medial tibiofemoral were observed only if there was evidence of
maceration of the medial meniscus. The authors suggested that meniscus
maceration was a prerequisite for a large BML and that BMLs may be a con-
sequence of impact forces in a knee with aberrant load distribution or insta-
bility resulting from meniscus damage. A longitudinal study will be needed
to establish causality, however. As a further indication of the strong associ-
ation between BMLs and mechanical stress, BMLs are much more prevalent
in the medial tibiofemoral compartment of varus knees than in knees with
neutral or valgus alignment (approximately 74% versus 16%). In valgus
knees, BMLs tend to localize to the lateral compartment [85].

Relief of osteoarthritic pain by physiologic interventions: further reasons

to consider a mechanical etiopathogenesis of osteoarthritis
In considering risk factors for OA from an epidemiologic perspective,
Felson [91] suggested that elimination of obesity, prevention of knee injury,
and elimination of jobs requiring knee bending and carrying heavy loads, all
of which can cause increased stress on the joints under consideration, would
decrease the incidence of OA of the hip and knee significantly. The authors
of this article contend that such considerations may be important both in the
prevention of incident OA and in the treatment of established symptomatic
disease. Several examples of the relief of OA pain by a mechanical (ie, phys-
iologic) rather than pharmacologic intervention can be cited.
For example, patients who have with knee OA who undergo valgus
osteotomy to alter the mechanical stresses across the hip joint report relief
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 551

of knee pain on follow-up examination. Notably, improvement at 2 years

was unrelated to whether the hyaline articular cartilage remains unchanged
from baseline, shows worsening of OA pathology, or is replaced with fibro-
cartilage [92]. Langlais and colleagues [93] reported that 83% of patients
who had severe hip OA were pain free 3 to 10 years (average, 6 years) after
undergoing valgus osteotomy.
Reduction of the abnormally high stresses in an OA joint by osteotomy
has been shown to result in gradual loss of the subchondral sclerosis in the
radiograph (see Fig. 3) and an increase in the formerly narrowed joint
space [30,31]. As indicated earlier, whether the latter represents growth of
new hyaline cartilage or fibrocartilage may be immaterial to the clinical
outcome.
A recent systematic review and meta-analysis of patellar taping for patel-
lofemoral OA by Warden and colleagues [94] focuses on another example of
successful mechanical treatment of painful OA. Taping to exert a medially
directed force on the patella of patients who have symptomatic radiographic
patellofemoral OA significantly decreased chronic knee pain in comparison
with no taping or sham taping. The results are consistent with evidence that
anterior knee pain in patients who have patellofemoral OA is caused by
lateral displacement of the patella, with increased loading of the lateral facet
[95]. Surgical advancement of the tibial tubercle, to weaken the quadriceps
muscle by increasing its leverage, also can have long-term symptomatic
benefits in patients who have patellofemoral OA [96].
Recent preliminary reports of the results of joint distraction in the treat-
ment of severe symptomatic OA provide further evidence; for example,
among nine patients who had severe tibiofemoral OA who were candidates
for joint arthroplasty, distraction of the knee joint for 2 months via an
external fixation frame that was fitted with springs that bridged the knee
and hinges to maintain intermittent synovial fluid pressure resulted in strik-
ing reductions in joint pain and improvement in function and clinical status,
with almost complete normalization by 6 months [97]. In five patients who
have been followed for 2 years after undergoing knee distraction, the
improvement has been sustained. Longitudinal MRI examinations suggest
that articular cartilage repair occurs, with a 30% increase in total cartilage
volume and a 25% increase in mean cartilage thickness, relative to baseline.
Urinary biomarker analyses were consistent with high levels of turnover
of articular cartilage and bone during distraction, which subsequently nor-
malized. These results are consistent with demonstrations of the prolonged
clinical efficacy of distraction in the treatment of severe hip and ankle OA
[98–100] and with reports of clinical improvement in ankle OA persisting
for at least 7 years [101].
Earlier the authors pointed out the inverse relationship between knee
extensor strength and the risk of developing knee OA [36]. Several random-
ized, controlled trials have shown that quadriceps exercise regimens that
were diverse in content, intensity, and frequency improved quadriceps
552 BRANDT et al

strength, reduced joint pain, improved function, and reduced reflex inhibi-
tion of the quadriceps in patients who had knee OA [102]. As reviewed by
Minor [103], longer trials tended to be more effective than shorter ones;
the higher the dose of exercise, the more effective was the intervention;
and only weight-bearing exercise seemed to improve function.
Considerable interest has focused on the importance in knee OA of the
peak adduction moment, which is considered to be a proxy for the dynamic
load on the medial tibiofemoral compartment of the knee [104]. In cross-
sectional studies the peak adductor moment was significantly greater in
patients who had medial compartment knee OA than in controls and also
was significantly greater in patients who had more severe OA than in those
who had less severe disease [105,106]. Peak adductor moment has been
shown to be a strong predictor of radiographic progression in patients
who have medial compartment OA [107] and of the development of knee
pain in asymptomatic older subjects [108]. Chang and colleagues [109]
recently reported that a greater degree of toe-out during gait, which shifts
the ground reaction force vector closer to the center of the knee, thereby re-
ducing the adductor moment, decreased the risk of radiographic progression
of knee OA over 18 months.
In subjects who had relatively mild structural changes of knee OA,
Thorp and colleagues [110] found that those who had knee pain had signif-
icantly higher medial compartment loads than those who were asymptom-
atic; the loads in those who were asymptomatic were no different from
those in normal controls. The results suggest that at this stage of structural
damage, individuals who have symptomatic OA differ biomechanically
from those who have asymptomatic disease. The possibility that the two
groups differed in the prominence of repetitive impulsive loading warrants
consideration.
In gait analyses performed on subjects who had knee OA while they were
wearing their everyday walking shoes and also while walking barefoot, peak
loads at the hips and knees decreased significantly with barefoot walking,
with a nearly 12% reduction in the knee adduction moment [111]. In a recent
preliminary report [112], the same researchers found that a shoe that was
designed to incorporate essential features of natural foot motion reduced
dynamic loading of the medial compartment of the knee during gait.
Finally, Schnitzer and colleagues [113], in a 4-week study of 18 patients
who had symptomatic knee OA and varus deformity, found that treatment
with the NSAID piroxicam reduced the level of joint pain but significantly
increased the adductor moment; that is, it increased loading of the already
damaged medial tibiofemoral compartment. Whether this increase would
have resulted in acceleration of structural damage if treatment had been
maintained for longer periods (possibly because of a faster gait and greater
impulsive loading because of relief of joint pain) is not known. The question
has potential clinical importance, because many patients who have OA pain
are treated with NSAIDs for years.
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 553

Summary
The importance of abnormal joint mechanics in the etiopathogenesis of
OA cannot be overemphasized. The authors view OA as a process that is
attempting to contain a mechanical problem in the joint; OA is best defined
as failed repair of joint damage that has been caused by excessive mechan-
ical stress. Because the body’s innate mechanisms for repairing damaged
cartilage cannot deal with the underlying mechanical abnormality, they
cannot solve the problem of OA. Remodeling of the subchondral bone
may reduce the excessive levels of local stress and contain the mechanical
abnormality but also may cause joint pain and reactivate enchondral ossifi-
cation at the expense of the overlying articular cartilage.
The common mechanical factor underlying OA is a pathologic increase in
intra-articular stress, which can result either from a decrease in the load-
bearing area of the joint or a quantitative increase in, or aberration of, joint
loading (ie, repetitive impulsive loading). Impulsive loads cause microinjury
of both the subchondral bone and articular cartilage that may exceed the
ability of the joint to repair the damage.
Although the capacity for intrinsic repair of damaged articular cartilage is
limited, cells that are extrinsic to the cartilage can provide a mechanism for
repair if the local environment permits. The new cartilage they produce, fibro-
cartilage, is not histologically, biochemically, or biomechanically comparable
to normal hyaline articular cartilage, but it nonetheless can permit normal
joint function, prevent further deterioration, and, most importantly, permit
the patient to function asymptomatically. Neither the large role that the reduc-
tion of excessive levels of mechanical stress plays in promoting the healing
response in OA nor the evidence that the relief of joint pain and improvement
of functiondand not the appearance of the articular surfacedare the most
important outcomes of the healing process have been sufficiently emphasized.
On the other hand, excessive emphasis has been placed on the inflamma-
tory changes in OA that, although relevant to the patient’s symptoms, are
secondary to breakdown products of cartilage and bone, and on articular
cartilage as the major focus of OA damage and progression. It must be remem-
bered that the synovial joint is an organ and OA is organ failure that can be
initiated by abnormalities arising in any of its tissues. The authors believe
that the lack of focus on the loss of the habitually loaded contact area of
the joint and on aberrant loading have misdirected the therapeutic efforts
in OA.
How OA is viewed matters because of the implications for treatment
of patients. The authors attribute the lack of success in developing struc-
ture-modifying OA drugs to the fact that OA, from its earliest stages and
in whichever joint it occurs, is mechanically induced. A variety of mechanical
abnormalities can trigger the many processes involved in repair, and attempts
to contain the mechanical insult without a return to mechanical normality
will fail. Drugs cannot initiate this repair. The authors believe that it is very
554 BRANDT et al

unlikely that a drug that inhibits a specific enzyme or cytokine in the path-
ways of cartilage breakdown or that further stimulates the already increased
synthesis of cartilage matrix molecules by the chondrocytes will solve the
problem of OA as long as the joint remains in the same adverse mechanical
environment that got it into trouble in the first place. In addition, because
the subchondral bone is critical for containment of the mechanical abnormal-
ities that damage the cartilage, emphasis on cartilage repair alone is likely to
be futile. If the abnormal stresses on the joint are corrected, pharmacologic
intervention with a structure-modifying drug may be superfluous. Also, there
is no assurance that a drug that slowed the progression of structural damage
in an OA joint will have a beneficial effect on symptoms.
The authors suggest that the genes whose identification will be required
to prevent and treat cases of OA that do not follow significant (macro) joint
trauma will not be those that regulate chondrocyte metabolism but instead
those that control developmental and congenital joint deformities, that
reduce the area of the habitually loaded surface, or that underlie micro-
incoordination, resulting in the concentration of peak dynamic loads that
lead to cumulative tissue microdamage and to remodeling that is detrimen-
tal to joint function.

References
[1] Brandt KD, Mankin HJ, Shulman LE. Workshop on etiopathogenesis of osteoarthritis.
J Rheumatol 1986;13:1126–60.
[2] Kuettner K, Goldberg VM. Introduction. In: Kuettner K, Goldberg VM, editors. Osteoar-
thritic disorders. Rosemont (IL): American Academy of Orthopaedic Surgeons; 1995.
p. xxi–v.
[3] Knowlton RG, Katzenstein PL, Moskowitz RW, et al. Genetic linkage of a polymorphism
in the type II procollagen gene (COL 2A1) to primary osteoarthritis associated with mild
chondrodysplasia. N Engl J Med 1990;322:526–30.
[4] Radin EL, Burr DB, Caterson B, et al. Mechanical determinants of osteoarthrosis. Semin
Arthritis Rheum 1991;21(Suppl 2):12–21.
[5] Radin EL, Burr DB. Hypothesis: joints can heal. Sem Arthr Rheum 1984;13:293–302.
[6] Radin EL, Yang KH, Riegger C, et al. Relationship between lower limb dynamics and knee
joint pain. J Orthop Res 1991;9:398–405.
[7] Radin EL, Paul IL, Pollock D. Animal joint behavior under excessive loading. Nature 1970;
226:554–5.
[8] Convery FR, Akeson WH, Keown GH. The repair of large osteochondral defects. An
experimental study in horses. Clin Orthop Relat Res 1972;82:253–62.
[9] Palmoski M, Brandt KD. Immobilization of the knee prevents osteoarthritis after anterior
cruciate ligament transaction. Arthritis Rheum 1982;25:1201–8.
[10] Brandt KD, Radin EL, Dieppe PA, et al. Yet more evidence that OA is not a cartilage
disease. Ann Rheum Dis 2006;65:1261–4.
[11] Doherty M, Watt I, Dieppe P. Influence of primary generalized osteoarthritis on develop-
ment of secondary osteoarthritis. Lancet 1983;2:8–11.
[12] Schouten JSAG, van den Ouweland FA, Valkenburg HA. A twelve-year follow-up study in
the general population on prognostic factors of cartilage loss in osteoarthritis of the knee.
Ann Rheum Dis 1992;51:932–7.
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 555

[13] Smythe HA. The mechanical pathogenesis of generalized osteoarthritis. J Rheumatol 1983;
10(Suppl 9):10–2.
[14] American College of Rheumatology clinical slide collection. Atlanta (GA): American
College of Rheumatology; 1972–2004.
[15] Hernborg J, Nilsson BE. The relationship between osteophytes in the knee joints, osteoar-
thritis and aging. Acta Orthop Scand 1973;44:69–74.
[16] Radin EL. Osteoarthrosis: a perspective. Acta Orthop Scand 1995;66(Suppl 266):6–9.
[17] Weinstein SL. Long-term follow-up of pediatric orthopaedic conditions. Natural history
and outcomes of treatment. J Bone Joint Surg Am 2000;82:980–90.
[18] Fitzgerald JB, Jin M, Dean D, et al. Mechanical compression of cartilage explants induces
multiple time-dependent gene expression patterns and involves intracellular calcium and
cyclic AMP. J Biol Chem 2004;279:19502–11.
[19] Radin EL, Boyd RD, Martin RB, et al. Mechanical factors influencing cartilage damage.
In: Peyron JG, editor. Osteoarthritis: current clinical and fundamental problems. Paris:
Geigy; 1985. p. 90–9.
[20] Jones CM, Watt DG. Muscular control of landing from unexpected falls in man. J Physiol
1971;219:729–37.
[21] Lee JH, Fitzgerald JB, Dimmico MA, et al. Mechanical injury of cartilage explants causes
specific time-dependant changes in chondrocyte gene expression. Arthritis Rheum 2005;52:
2386–95.
[22] Lee JH, Bai Y, Flannery CR, et al. Cartilage mechanical injury and co-culture with joint
capsule tissue increase abundance of ADAMTS-5 protein and aggrecan-G1-NITEGE
product. Transactions of the Orthopedic Research Society 2006;31.
[23] Stokes M, Young A. The contribution of reflex inhibition to arthrogenous muscle weak-
ness. Clin Sci 1984;67:7–14.
[24] Felson DT, Kim Y. The futility of current approaches to chondroprotection. Arthritis
Rheum 2007;56:1378–83.
[25] Brandt KD, Radin EL, Dieppe PA, et al. The futility of current approaches to chondropro-
tectionda different perspective: comment on the special article by Felson and Kim [letter].
Arthritis Rheum 2007;56:3873–4.
[26] Myers SL, Flusser D, Brandt KD, et al. Prevalence of cartilage shards in synovium and their
association with synovitis in patients with early and end stage osteoarthritis. J Rheumatol
1992;19:1247–51.
[27] Burr DR, Schaffler MB. The involvement of subchondral mineralized tissues in osteoarth-
rosis: quantitative microscopic evidence. Microsc Res Tech 1997;37:353–7.
[28] Burr DB, Radin EL. Microfractures and microcracks in subchondral bone: are they rele-
vant to osteoarthrosis. Rheum Dis Clin North Am 2003;29:675–85.
[29] Meachim G, Bentley G. Horizontal splitting in patellar articular cartilage. Arthritis Rheum
1978;21:669–74.
[30] Pauwels F. Biomechanics of the normal and diseased hip: theoretical foundation, technique
and results of treatment. Berlin: Springer-Verlag; 1976. p. 129–268.
[31] Maquet P, Radin EL. Osteotomy as an alternative to total hip replacement in young adults.
Clin Orthop 1977;123:138–42.
[32] Hill AV. Production and absorption of work by muscle. Science 1960;131:897–903.
[33] Radin EL. Role of muscles in protecting athletes from injury. In: Astrand PO, Grimby G,
editors. Physical activity in health and disease. Acta Med Scand 1986;2(Suppl 711):143–7.
[34] Radin EL, Yang KH, Whittle MW, et al. The generation and transmission of heelstrike tran-
sient and the effect of quadriceps paralysis. In: Harris D, editor. Gait analysis and medical
photogrammetry. Oxford (UK): Oxford Orthopaedic Engineering Centre; 1987. p. 34–45.
[35] Hurley MV. The role of muscle weakness in the pathogenesis of osteoarthritis. Rheum Dis
Clin North Am 1999;25:299–314.
[36] Slemenda C, Brandt KD, Heilman D, et al. Quadriceps weakness and osteoarthritis of the
knee. Ann Intern Med 1997;127:97–104.
556 BRANDT et al

[37] Slemenda C, Heilman D, Brandt KD, et al. Reduced quadriceps strength relative to body
weight: a risk factor for knee osteoarthritis in women? Arthritis Rheum 1998;41:1951–9.
[38] Brandt KD, Heilman DK, Slemenda C, et al. Quadriceps strength in women with radio-
graphically progressive osteoarthritis of the knee and those with stable radiographic
changes. J Rheumatol 1999;26:2431–7.
[39] Radin EL, Paul IL. Does cartilage compliance reduce skeletal impact loads? The relative
force-attenuating properties of articular cartilage, synovial fluid, periarticular soft tissues
and bone. Arthritis Rheum 1970;13:139–44.
[40] Burr DB. Subchondral bone in the pathogenesis of osteoarthritis. Mechanical aspects. In:
Brandt KD, Doherty M, Lohmander LS, editors. Osteoarthritis. 2nd edition. Oxford (UK):
Oxford University Press; 2003. p. 125–33.
[41] Hayes WC, Swenson LW Jr, Schurman DJ. Axisymmetric finite element analysis of the
lateral tibial plateau. J Biomech 1978;11:21–33.
[42] Mente P, Lewis JL. The elastic modulus of calcified cartilage is an order of magnitude less
than that of subchondral bone. J Orthop Res 1994;12:637–47.
[43] Redler I, Mow VC, Zimny ML, et al. The ultrastructure and biomechanical significance of
the tidemark of articular cartilage. Clin Orthop 1975;112:357–62.
[44] Radin EL, Paul IL. A consolidated concept of joint lubrication. J Bone Joint Surg Am 1972;
54:607–16.
[45] Finlay JB, Repo RU. Cartilage impact in vitro: effect of bone and cement. J Biomech 1978;
11:379–88.
[46] Radin EL, Paul IL, Lowy M. A comparison of the dynamic force transmitting properties of
subchondral bone and articular cartilage. J Bone Joint Surg Am 1970;52:444–56.
[47] Hoshino A, Wallace WA. Impact-absorbing properties of the human knee 1987. J Bone
Joint Surg 1987;69B:807–11.
[48] Burr DB. The importance of subchondral bone in the progression of osteoarthritis. J Rheu-
matol 2004;31(Suppl 70):77–80.
[49] Li B, Aspden RM. Mechanical and material properties of the subchondral bone plate from
the femoral head of patients with osteoarthritis or osteoporosis. Ann Rheum Dis 1997;56:
247–54.
[50] Grynpas MD, Alpert B, Katz I, et al. Subchondral bone in osteoarthritis. Calcif Tissue Int
1991;49:20–6.
[51] Brown TD, Radin EL, Martin RB, et al. Finite element studies of some juxtarticular stress
changes due to localized subchondral stiffening. J Biomech 1984;17:11–24.
[52] Vener MJ, Thompson RC Jr, Lewis JL, et al. Subchondral damage after acute transartic-
ular loading: an in vitro model of joint injury. J Orthop Res 1992;10:759–65.
[53] McCrae F, Shouls J, Dieppe P, et al. Scintigraphic assessment of osteoarthritis of the knee
joint. Ann Rheum Dis 1992;51:938–42.
[54] Franchi A, Bullough PG. Secondary avascular necrosis in coxarthrosis: a morphologic
study. J Rheumatol 1992;19:1263–8.
[55] Newberry WN, Zukosky OK, Haut RC. Subfracture insult to a knee joint causes alter-
ations in the bone and in the functional stiffness of overlying cartilage. J Orthop Res
1997;15:450–5.
[56] Dedrick OK, Goulet R, Huston L, et al. Early bone changes in experimental osteoarthritis
using microscopic computer tomography. J Rheumatol 1991;18(Suppl 27):44–5.
[57] Dedrick OK, Goldstein SA, Brandt KD, et al. A longitudinal study of subchondral plate and
trabecular bone in cruciate-deficient dogs with osteoarthritis followed for up to 54 months.
Arthritis Rheum 1993;36:1460–7.
[58] Myers SL. Effects of a bisphosphonate on bone histomorphometry and dynamics in the
canine cruciate-deficiency model of osteoarthritis. J Rheumatol 1999;26:2546–53.
[59] Liang MH. Pushing the limits of patient-oriented outcome measurements in the search for
disease modifying treatments for osteoarthritis. J Rheumatol Suppl 2004;31:61–5.
[60] Hadler NM. Knee pain is the maladydnot osteoarthritis. Ann Intern Med 1992;116:598–9.
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 557

[61] Massardo L, Watt I, Cushnaghan J, et al. Osteoarthritis of the knee joint: an eight year
prospective study. Ann Rheum Dis 1989;48:893–7.
[62] Lawrence JS, Bremmer JM, Bier F. Osteo-arthrosis. Prevalence in the population and
relationship between symptoms and x-ray changes. Ann Rheum Dis 1966;25:1–24.
[63] Lawrence JS. Osteoarthrosis. In: Lawrence JS, editor. Rheumatism in populations.
London: Heinemann Medical Books, Ltd.; 1977. p. 98–155.
[64] Felson DT, Naimark A, Anderson J, et al. The prevalence of knee osteoarthritis in the
elderly. Arthritis Rheum 1987;30:914–8.
[65] Sandell LS, Heinegard D, Hering TM. Cell biology, biochemistry, and molecular biology of
articular cartilage in osteoarthritis. In: Moskowitz RW, Altman RD, Hochberg MC, et al,
editors. Osteoarthritis. 4th edition. Philadelphia: Lippincott Willams and Wilkins; 2007.
p. 73–106.
[66] Poole AR, Guilak F, Abramson SB. Etiopathogenesis of osteoarthritis. In: Moskowitz
RW, Altman RD, Hochberg MC, et al, editors. Osteoarthritis. 4th edition. Philadelphia:
Lippincott Williams and Wilkins; 2007. p. 27–49.
[67] Sandy J. Proteolytic degradation of normal and osteoarthritic cartilage matrix. In: Brandt
KD, Doherty M, Lohmander LS, editors. Osteoarthritis. 2nd edition. Oxford (UK): Oxford
University Press; 2003. p. 82–92.
[68] O’Reilly SC, Muir KR, Doherty M. Knee pain and disability in the Nottingham commu-
nity: association with poor health status and psychological distress. Br J Rheumatol
1998;57:588–94.
[69] Evans CH, Mears DC, McKnight JL. A preliminary ferrographic survey of the wear
particles in human synovial fluid. Arthritis Rheum 1981;24:912–8.
[70] Evans CH. Cellular mechanisms of hydrolytic enzyme release in osteoarthritis. Semin
Arthritis Rheum 1981;11(Suppl 1):93–5.
[71] Boniface RJ, Cain PR, Evans CH. Articular responses to purified cartilage proteoglycans.
Arthritis Rheum 1998;31:258–66.
[72] Schumacher HR, Gordon G, Paul H, et al. Osteoarthritis, crystal deposition and inflamma-
tion. Semin Arthritis Rheum 1981;11:116–9.
[73] Jasin HE. Immune mechanisms in osteoarthritis. Semin Arthritis Rheum 1989;18(Suppl 2):
86–90.
[74] Myers SL, Brandt KD, Ehlich JW, et al. Synovial inflammation in patients with early oste-
oarthritis of the knee. J Rheumatol 1990;17:1662–9.
[75] Hill CL, Gale DG, Chaisson CE, et al. Knee effusions, popliteal cysts and synovial thick-
ening. Association with knee pain in those with and without osteoarthritis. J Rheumatol
2001;28:1330–7.
[76] Hill CL, Hunter DJ, Niu J, et al. Synovitis detected on magnetic resonance imaging and its
relation to pain and cartilage loss in knee osteoarthritis. Ann Rheum Dis 2007;66:1599–603.
[77] Lo G, McAlindon T, Niu J, et al. Strong association of bone marrow lesions and effusion
with pain in osteoarthritis. Arthritis Rheum 2007;56(Suppl):S790.
[78] Arnoldi CC, Linderholm H, Mussbichler H. Venous engorgement and intraosseous hyper-
tension in osteoarthritis of the hip. J Bone Joint Surg 1972;54B:409–21.
[79] Arnoldi CC, Lempberg RK, Linderholm H. Immediate effect of osteotomy on the intrame-
dullary pressure of the femoral head and neck in patients with degenerative osteoarthritis.
Acta Orthop Scand 1971;42:357–65.
[80] Arnoldi CC, Djurhuus JC, Heerfordt J, et al. Intraosseous phlebography, intraosseous
pressure measurements and 99m Tc-polyphosphate scintigraphy in patients with various
painful conditions in the hip and knee. Acta Orthop Scand 1980;51:19–28.
[81] McAlindon TEM, Watt I, McCrae F, et al. Magnetic resonance imaging in osteoarthritis of
the knee: correlation with radiographic and scintigraphic findings. Ann Rheum Dis 1991;
50:14–9.
[82] Dieppe P, Cushnaghan J, Young P, et al. Prediction of the progression of joint space narrowing
in osteoarthritis of the knee with bone scintigraphy. Ann Rheum Dis 1993;52:557–63.
558 BRANDT et al

[83] Hunter DJ, Niu J, Zhang Y, et al. Altered perfusion and venous hypertension is present in
regions of bone affected by BMLs in knee OA. Osteoarthritis Cartilage 2007;15(Suppl C):
C171.
[84] Zanetti M, Bruder E, Romero J, et al. Bone marrow edema pattern in osteoarthritic knees:
correlation between MR imaging and histologic findings. Radiology 2000;215:835–40.
[85] Felson DT, McLaughlin S, Goggins J, et al. Bone marrow edema and its relation to progres-
sion of knee osteoarthritis. Ann Intern Med 2003;139:330–6.
[86] Felson DT, Chaisson CE, Hill CL, et al. The association of bone marrow lesions with pain
in knee osteoarthritis. Ann Intern Med 2001;134:541–9.
[87] Meizer R, Radda C, Stolz G, et al. MRI-controlled analysis of 104 patients with painful
bone marrow edema in different joint locations treated with the prostacyclin analogue
iloprost. Wien Klin Wochenschr 2005;8:278–86.
[88] Aigner N, Petje G, Schneider W, et al. Bone marrow edema syndrome of the femoral head:
treatment with the prostacyclin analogue iloprost vs. core decompression: an MRI-
controlled study. Wien Klin Wochenschr 2005;117:130–5.
[89] Felson DT, Niu J, Guermazi A, et al. Correlation of the development of knee pain with
enlarging bone marrow lesions on magnetic resonance imaging. Arthritis Rheum 2007;
56:2986–92.
[90] Lo G, Hunter D, Nevitt M, et al. Strong association of meniscal maceration and bone
marrow lesions in osteoarthritis. Arthritis Rheum 2007;56(Suppl):S125.
[91] Felson DT. Preventing knee and hip osteoarthritis. Bull Rheum Dis 1998;47:1–4.
[92] Bergennud H, Johnell O, Redlund-Johnell I, et al. The articular cartilage after osteotomy
for gonarthrosis: biopsies after 2 years in 19 cases. Acta Orthop Scand 1992;63:413–6.
[93] Langlais F, Roure JL, Maquet P. Valgus osteotomy in severe osteoarthritis of the hip.
J Bone Joint Surg 1979;61B:424–31.
[94] Warden SJ, Hinman RS, Watson MA Jr, et al. Patellar taping and bracing for the treatment
of chronic knee pain: a systematic review and meta-analysis. Arthritis Rheum 2008;59:73–83.
[95] Niu J, Zhang YQ, Nevitt M, et al. Patellar malalignment is associated with prevalent
patellofemoral osteoarthritis: the Beijing osteoarthritis study. Arthritis Rheum 2005;
52(Suppl 9):S456–7.
[96] Maquet P. Biomechanics of the knee. With application to the pathogenesis and the surgical
treatment of osteoarthritis. Berlin: Springer; 1984. p. 144–56; 279–82.
[97] Intema F, van Roermund PM, Castelein RM, et al. Joint distraction in the treatment of knee
osteoarthritis; the first clinical results. Osteoarthritis Cartilage 2007;15(Suppl C):C234.
[98] Aldegheri R, Trivella G, Saleh M. Articulated distraction of the hip: conservative surgery
for arthritis in young patients. Clin Orthop Relat Res 1994;301:94–101.
[99] Marijnissen AC, Van Roermund PM, van Melkebeek J, et al. Arthritis Rheum 2002;46:
2893–902.
[100] Paley D, Lamm BM. Ankle joint distraction. Foot Ankle Clin 2005;10:685–98.
[101] Ploegmakers JJ, van Roermund PM, van Melkebeek J, et al. Prolonged clinical benefit from
joint distraction in the treatment of ankle osteoarthritis. Osteoarthritis Cartilage 2005;13:
582–8.
[102] Minor MA. Exercise for the patient with osteoarthritis. In: Brandt KD, Doherty M,
Lohmander LS, editors. Osteoarthritis. 2nd edition. Oxford (UK): Oxford University
Press; 2003. p. 299–310.
[103] Minor MA. Impact of exercise on osteoarthritis outcomes. J Rheumatol 2004;31(Suppl 70):
81–6.
[104] Birmingham TB, Hunt MA, Jones IC, et al. Test-retest reliability of the peak knee adduc-
tion moment during walking in patients with medial compartment knee osteoarthritis.
Arthritis Rheum 2007;57:1012–7.
[105] Baliunas AJ, Hurwitz DE, Ryals AB, et al. Increased knee joint loads during walking are
present in subjects with knee osteoarthritis. Osteoarthritis Cartilage 2002;10:573–9.
 ETIOPATHOGENESIS OF OSTEOARTHRITIS 559

[106] Mundermann A, Dyrby CO, Hurwitz DE, et al. Potential strategies to reduce medial com-
partment loading in patients with knee osteoarthritis of varying severity: reduced walking
speed. Arthritis Rheum 2004;50:11172–8.
[107] Miyazaki T, Wada M, Kawahara H, et al. Dynamic load at baseline can predict radio-
graphic disease progression in medial compartment knee osteoarthritis. Ann Rheum Dis
2002;61:617–22.
[108] Amin S, Luepongsak N, McGibbon CA, et al. Knee adduction moment and development
of chronic knee pain in elders. Arthritis Rheum 2004;51:371–6.
[109] Chang A, Hurwitz D, Dunlop D, et al. The relationship between toe-out angle during gait
and progression of medial tibiofemoral osteoarthritis. Ann Rheum Dis 2007;66:1271–5.
[110] Thorp LE, Sumner DR, Wimmer MA, et al. Relationship between pain and medial knee
joint loading in mild radiographic knee osteoarthritis. Arthritis Rheum 2007;57:1254–60.
[111] Shakoor N, Block JA. Walking barefoot decreases loading on the lower extremity joints in
knee osteoarthritis. Arthritis Rheum 2006;54:2923–7.
[112] Shakoor N, Lidtke RH, Sengupta M, et al. Mobility footwear reduces dynamic loads in
subjects with osteoarthritis of the knee. Osteoarthritis Cartilage 2007;15:C219.
[113] Schnitzer TJ, Popovich JM, Andersson GB, et al. Effect of piroxicam on gait in patients
with osteoarthritis of the knee. Arthritis Rheum 1993;36:1207–13.