Review

Osteoarthritis in women: effects of

estrogen, obesity and physical activity
Jennifer E Stevens-Lapsley†1 & Wendy M Kohrt2
Arthritis is the leading cause of physical disability in the USA and osteoarthritis (OA), the most
common form, affects nearly 27 million people. Women have a higher prevalence of OA than men,
but the underlying causes for the increased susceptibility of women to OA are not fully understood.
This review discusses the associations of sex hormones, obesity and physical activity with the incidence
and progression of OA in women. Although many studies have explored the relationships of estrogen
and reproductive history with the risk of OA or joint replacement, there is no consensus on the role
of these factors. Obesity clearly increases risk for OA, but whether obesity affects women and men
differently is uncertain. Moderate levels of physical activity do not appear to increase the incidence
or progression of OA and may even have a weak protective effect. Future investigations should focus
on sex-specific mechanisms for the development of OA and determine whether there are modifiable
factors (e.g., estrogen, obesity, physical activity) that can be targeted through prevention and
treatment strategies to mitigate the increased incidence and prevalence of OA in women.

Background on osteoarthritis Although the precise etiology is unknown, OA

Arthritis is the leading cause of physical dis- probably results from a combination of metabolic,
ability in the USA. The most common form of genetic and mechanical factors (Box 1) [7–11] . Some
arthritis is osteoarthritis (OA), which affects of the prominent risk factors for the development
nearly 27  million people in the USA  [1] . OA of OA include older age, female sex, obesity, pre-
causes deterioration of the joint cartilage and vious knee trauma, occupational load, ethnicity
formation of reactive new bone at the margins and genetics [4,10,12,13] . Other risk factors for OA
and subchondral areas of the joints. This degen- incidence that are more controversial include sex
eration results from a breakdown of chondro­ hormones (described further later), biomechanical
cyte function, most commonly in the distal malalignment, joint laxity and quadriceps muscle
and proximal interphalangeal joints, but also weakness [7,9,14–16] . It has been suggested that the
in the hip and knee joints [2,3] . More recently, incidence and progression of symptomatic and
it has become apparent that OA affects not only radiographic knee OA involve different processes
articular cartilage, but also the entire joint. In [17,18] . If true, risk factors for the progression of
addition to erosion of articular cartilage, there OA may be different than those for the devel-
are alterations in subchondral bone, degenera- opment OA. A systematic review of prognos- 1
University of Colorado – Anschutz
tion of menisci, inflammation of the synovium, tic f­actors for progression of OA found strong Medical Campus, Physical Therapy
Program, Department of Physical
and bone and cartilage overgrowth (i.e., osteo- evidence that both higher serum hyaluronic Medicine & Rehabilitation, Aurora,
phytes) [4] . More detail regarding the patho- acid levels and generalized OA were predictive CO 80045, USA
physiology of OA has been reviewed recently of knee OA progression [19] , whereas sex, knee 2
University of Colorado – Anschutz
Medical Campus, Division of Geriatric
by Hunter [4] . pain, radiologic severity, knee injury, quadriceps Medicine, Department of Medicine
Clinical signs of symptomatic OA include strength and regular sport activities were not [19] . †
Author for correspondence:
Tel.: +1 303 724 9170
slowly developing joint pain, stiffness and The review also found conflicting evidence as to Fax: +1 303 724 9016
enlargement with limitation of motion. whether body mass index (i.e., o­besity) and age jennifer.stevens-lapsley@ucdenver.edu
Radiographic signs include osteophytes, joint influence the progression of OA [19] .
space narrowing and sclerosis [3] . Pain symp- The factors that influence the development
toms are not always present with OA, even when and progression of OA are not fully understood. Keywords
extensive radiographic damage is evident [5–7] . The intent of this review was to discuss whether
• estrogen • hormone • obesity
Felson and others reported that only 40–80% sex hormones, obesity and physical activity • osteoarthritis • physical activity
of people with radiographic evidence of OA influence the incidence and progression of OA • sex • women
have clinical symptoms [6,7] . Radiographic OA in a sex-specific manner. After careful review of
of the knee is present in more than 33% of the literature, we have chosen to emphasize some
people over 60 years of age, whereas 10–15% of studies over others in an effort to summarize part of
people over 60 years of age have symptomatic findings, while providing a fair and balanced
knee OA, defined as knee pain on most days [7] . summary of the existing literature.

10.2217/WHE.10.38 © 2010 Future Medicine Ltd Women's Health (2010) 6(4), 601–615 ISSN 1745-5057 601
REVIEW – Stevens-Lapsley & Kohrt

Above the age of 50 years (which approximates

Box 1. Known and suspected risk
the age of menopause), the incidence of OA rises
factors for developing osteoarthritis.
more steeply in women than in men [27] , which
Known makes it tempting to postulate that ovarian
• Age hormones play a protective role in joint health
• Female sex [3,14,28,29] . The observations that articular chon-
• Obesity drocytes have functional estrogen receptors and
• Previous knee trauma
that estrogen can upregulate proteoglycan syn-
• Repeated overuse, especially occupational
thesis [30,31] , bolster the scientific plausibility for
• Knee bending or lifting (knee osteoarthritis)
direct benefits of estrogens on articular cartilage.
• Genetics
• Ethnicity
However, although many studies have exam-
ined female sex hormones as determinants of
Suspected or unconfirmed
• Estrogen therapy (protective?)
OA risk, there is no clear consensus opinion on
• Early hysterectomy/surgical menopause their mechanistic role [31,32] . Recent systematic
• Osteoporosis (protective) reviews evaluated multiple aspects of hormone
• Biomechanical malalignment exposure as potential determinants of hip, knee
• Muscle weakness and hand OA, including endogenous hormone
levels, age at menarche and menopause, dura-
Associations of OA with sex tion of fertile period, menopausal status, years
hormone factors since menopause, and oophorectomy [14,28] ; some
Women have a higher prevalence of OA than of these factors will be discussed further below.
men, especially in the knee (Table 1) [20–22] . OA The main conclusion was that the postulated
not only occurs more frequently in women than beneficial effects of ovarian hormones on joint
in men, but also with greater severity [23,24] . health were not supported by the available evi-
Women often have more joints affected and dence. Much of the evidence pointed toward no
report greater morning stiffness and joint swell- association between OA and exposure to female
ing than men [24] . Women waiting for knee or hip sex hormones.
replacement typically report higher levels of pain One of the largest investigations of the asso-
than men [25,26] . An investigation of the relation- ciations of estrogens and reproductive history
ship between symptoms and radiographic grades with advanced OA (i.e., joint replacement) was
of knee OA recently suggested that women have conducted as part of the Million Women Study
more severe symptoms and symptom progression through the National Health Service breast
than men presenting with the same radiographic screening centers in England and Scotland [29] .
grade of OA [23] . Information was collected from 1,306,081

Table 1. Prevalence of knee osteoarthritis.

Demographic subgroup Radiographic knee OA (%) Symptomatic knee OA (%)
Male (age in years) 31.2 10.0
60–69 27.4 7.5
70–79 33.5 12.9
≥80 40.7 12.7
Female (age in years) 42.1 13.6
60–69 35.2 10.6
70–79 44.6 15.3
≥80 55.6 18.2
BMI (kg/m ) 2

<25 23.8 6.5

25–29 36.9 9.9
≥30 57.4 23.2
Although OA affects other joints similarly, the largest percentage of OA occurs in the knee; therefore prevalence data are
presented for the knee by sex, age and BMI.
OA: Osteoarthritis.
Data taken from [20].

602 www.futuremedicine.com future science group

 Osteoarthritis in women: effects of estrogen, obesity & physical activity – Review

women at study entry on age at menarche, par- all [29,40] , found a lower incidence or prevalence
ity, age at menopause, menopausal status, and of knee and hip OA in HT users, supporting the
use of sex hormone therapies. Women were fol- concept that estrogens have a protective effect on
lowed an average of 6.1 years, over which time joints. In the prospective Chingford study, 715
the incidence rates of total knee and total hip middle-aged women were followed for 4 years to
replacement were evaluated. The major findings determine the incidence of knee OA, defined by
from this large prospective cohort study were: the development of osteophytes and joint space
• There was a linear trend for increased parity narrowing, and whether age, weight and HT are
to be asso­ciated with increased risk for joint risk factors for OA [34] . Age and obesity were the
replacement, and the effect was more pro- strongest determinants of the incidence of knee
nounced for the hip than the knee. The rela- OA; current HT use conferred a nonsignificant,
tive risks (RR) for hip and knee replacement protective effect (overall response [OR]: 0.41;
increased by 2 and 8%, respectively, per birth; 95% CI: 0.12–1.42) [34] .
The Framingham Osteoarthritis Study found
• Early age of menarche (aged 11  years or that postmenopausal women who used HT had
younger) was associated with an increased RR lower odds of developing radiographic knee
of hip (RR: 1.09; 95% CI: 1.03–1.16) and OA and the protective effect was greater with
knee (RR: 1.15; 95% CI: 1.08–1.22) increased duration of HT treatment [33,38] . In this
re­placement; study, 551 women were categorized according to
• When compared with women who had never- HT use (349 never-users; 162 past users; 40 cur-
used menopausal hormone therapy (HT), cur- rent users) [38] . Over 8 years of follow-up, 17.4%
rent use was associated with an increased RR of never-users had knee radiographic scores that
of hip (RR: 1.38; 95% CI: 1.30–1.46) and worsened by 1 grade and 5.8% by 2 or 3 grades.
knee replacement (RR: 1.58; 95% CI: 1.48– Among current HT users, only 11.7% of knee
1.69). RR was lower in past users of meno- radiographic scores worsened by 1 grade and none
pausal HT, but still significantly elevated (hip worsened by more than 1 grade. With adjust-
RR: 1.13; 95% CI: 1.06–1.21; knee RR: 1.39; ments for age and other potential confounding
95% CI: 1.29–1.49). Both estrogen-only (hip factors, the RR of incident radiographic knee
RR: 1.21; 95% CI: 1.10–1.33; knee RR: 1.48; OA, in comparison with never-users of HT,
95% CI: 1.35–1.63) and estrogen plus proges- was 0.8 (95% CI: 0.5–1.4) in past users and 0.4
tin therapies (hip RR: 1.51; 95% CI: 1.41– (95% CI: 0.1–3.0) in current users. Current use of
1.62; knee RR: 1.68; 95% CI: 1.54–1.83) were HT also showed a nonsignificant decrease in risk
associated with increased risk for joint replace- of progressive knee OA compared with never-use
ment. Past use of hormone contraceptive (OR: 0.5; 95% CI: 0.1–2.9). When both incident
t­herapy was not associated with increased risk. and progressive radiographic knee OA cases were
combined, current HT use was associated with
From a very broad perspective, this Million a nonsignificant 60% decreased risk compared
Woman Study suggested that increased exposure with never-use (OR: 0.4; 95% CI: 0.1–1.5). These
to sex hormones is associated with increased risk findings are seemingly in direct contrast with
for joint replacement, but alternative interpreta- the Million Women Study (discussed above),
tions were also discussed. For example, although which found that the use of menopausal HT was
the association of parity with joint replacement asso­ciated with an increased incidence of joint
risk was adjusted for BMI at the time of study r­eplacement [29] .
entry, the authors acknowledged that the asso- Few randomized controlled trials (RCTs)
ciation could have been influenced by increased have evaluated the effects of HT on OA from
weight-bearing during pregnancy, rather than by an intervention perspective, and none have
the sex hormone environment. They also sug- evaluated OA risk as a primary outcome [41,42] .
gested that the associations of HT with joint In the intention-to-treat analyses of data from
replacement risk could have reflected nonbio- the Women’s Health Initiative (WHI) interven-
logical factors, because women on therapy may tion trials, women randomized to estrogen-only
have had better access to healthcare services. therapy had a lower rate of joint replacement
than women on placebo therapy (hazard ratio
Estrogen-based hormone therapy [HR]: 0.84; 95% CI: 0.70–1.00); there were
Many epidemiological studies have investigated trends for benefits for both hip (HR: 0.73; 95%
the association of menopausal estrogen-based CI: 0.52–1.03) and knee replacement (HR: 0.87;
HT with OA. Some studies [33–39] , but not 95% CI: 0.71–1.07) [41] . When censored for

future science group Women's Health (2010) 6(4) 603

REVIEW – Stevens-Lapsley & Kohrt

nonadherence to assigned HT, women on of the use or duration of use with the develop-
estrogen therapy had significantly lower risks ment of OA [28] . The Million Women Study also
for any joint replacement (HR: 0.73; 95% CI: found that previous oral contraceptive use was
0.58–0.93) and hip replacement (HR: 0.55; not associated with the incidence of hip or knee
95% CI: 0.35–0.88), but not knee replacement replacement (RR: 1.02; 95% CI: 0.98–1.06; RR:
(HR: 0.80; 95% CI: 0.61–1.05). By contrast, 1.00; 95% CI: 0.96–1.04, respectively) [29] . Thus,
women randomized to estrogen plus progestin the available evidence suggests that use of hor-
therapy in the WHI trial did not have a change monal contraception does not influence the risk
in incidence of joint replacement (HR:  0.99; of developing OA.
95% CI: 0.82–1.20), and censoring for adher-
ence to therapy did not alter the results [41] . These Age of menopause
findings suggest that postmenopausal estrogen The increased incidence of OA around the time
therapy has beneficial effects on joint health, but of menopause suggests that the development of
that estrogen plus progestin therapy does not. OA is linked with the menopausal withdrawal
The Heart and Estrogen/Progestin Replacement of ovarian hormones [3] . However, attempts to
Study (HERS) also concluded that there was establish a temporal relationship between the
no significant effect of estrogen plus progestin onset of OA and menopause have been inconclu-
therapy on knee pain or related d­isability when sive. Several studies have found that neither age
compared with placebo therapy [42] . at menopause nor menopausal status are indepen-
A number of factors may explain the lack of dently related to incidence of OA [14,28,29,39,43] .
consensus for the effects of HT on OA. The By contrast, a large epidemiological study in Italy
highest level of evidence (i.e., RCTs) is lack- found that women experiencing surgical or natu-
ing, because only two large RCTs have been ral menopause were at increased risk of OA (OR:
performed and neither had joint replacement 1.13; 95% CI: 1.07–1.21; OR: 1.18; 95% CI:
or incident OA as a primary outcome [41,42] . 1.08–1.28) when compared with premenopausal
The prospective observational studies of inci- women, yet there was no clear relationship of age
dent OA in HT-users versus non-users may have at menopause with risk of OA  [39] . It has been
been influenced by confounding factors, such as postulated that the perimenopausal decrease in
access to healthcare services, different levels of progesterone levels results in a transient period of
physical activity or body weight [29,41,42] . It is also exposure to unopposed estrogens that increases
possible that HT has different effects on clinical risk of OA [44] . However, this hypothesis is seem-
knee symptoms than on the structural radio- ingly in contrast with the finding from the WHI
graphic evidence of OA [3,42] . Studies that have RCT that unopposed estrogen therapy had favor-
examined the association of HT with symptom- able effects on joint health. It remains unclear how
atic or clinical OA are less consistent than studies short-term hormonal imbalances or permanent
of radiographic OA in suggesting a protective declines in hormone concentrations contribute
role for HT [3,42] . HT may slow certain radio- to the surge in OA that appears to occur around
graphic changes of OA, but has little or no effect the age of menopause.
on symptoms. Investigations that focus on both
symptoms and radiographic evidence of OA will Age of menarche
be necessary to explore these potential differen- Only a few studies have investigated the rela-
tial effects. Finally, the WHI intervention trials tionship between age of menarche and OA
suggested that the beneficial effects of estrogens risk [45–47] . The Million Women study reported
on joint health are countered by progestins [41] . that a younger age at menarche (age 11 years or
Typically, prospective observational studies have less) was related to a greater risk of hip and knee
not examined the independent effects of effects replacement when compared with menarche at
of estrogen therapy versus estrogen plus pro- age 12  years (RR: 1.09; 95% CI: 1.03–1.16;
gestin therapy. Therefore, the type of HT and RR: 1.15; 95% CI: 1.08–1.22 for hip and knee
duration of use may confound inferences from replacement, respectively) [29] . Young age at
observational studies regarding the role of HT menarche was associated with increased risk for
in the incidence and progression of OA. hand OA in one study (n = 745) [46] , but not
in another (n = 348) [47] . However, the latter
Previous hormonal contraceptive use study did find that the total number of years of
A systematic review of the effects of hormonal menstruation was directly related to the pres-
contraception on OA of the hand, hip and knee ence of hand OA [47] . These findings suggest that
found that there was no significant association increased exposure to sex hormones promotes the

604 www.futuremedicine.com future science group

 Osteoarthritis in women: effects of estrogen, obesity & physical activity – Review

development of OA. However, because early age is that obesity is connected to an array of sys-
of menarche has been associated with increased temic and hormonal factors responsible for bone
risk for other age-related chronic diseases [48] , it and cartilage metabolism, especially in women
is possible that factors other than sex hormones [24,56,58] . Such factors could alter the density of
influence the association with OA. subchondral bone or exert metabolic actions
that have systemic effects on joints through-
Parity out the body [24] . For example, leptin, which is
Several studies have evaluated the effects of par- secreted primarily from adipocytes, may play an
ity and the risk for OA [8,29,36,40,43,47,49–51] . The important role in the onset and progression of
Million Women Study found that parity was OA [59] . Other adipocyte-derived factors, such
associated with a small (2–8%), but significant, as IL‑6 and C‑reactive protein, appear to be pro-
increase in RR of knee and hip joint replace- catabolic for chondrocytes [60] . Further research
ment [29] . Another study found increased risk of is necessary to evaluate whether leptin or other
hand OA with increasing number of births [47] , adipokines are important systemic or local fac-
suggesting that greater body weight during or tors to explain the link between obesity and OA.
after pregnancy does not fully account for the Furthermore, because women at any given BMI
increased OA risk in weight-bearing joints. have a greater total body adiposity (fat mass, per-
Nevertheless, despite the findings of these two cent body fat) than men [61] , adipocyte-derived
studies, the majority of investigations have systemic and hormonal factors may explain, in
found no relationship between parity and risk part, the higher incidence of OA in women.
for OA [8,36,40,43,49–51] . In light of the potential role of systemic fac-
tors influenced by obesity, it is difficult to iso-
Summary of associations of OA with sex late the biomechanical effects of BMI on the
hormone factors development and progression of OA. Many
Although a number of studies have examined obese patients exhibit abnormal lower extrem-
the relationships of sex hormone factors with the ity alignment (discussed later), which results in
risk of OA or joint replacement, there is currently increased joint-reactive forces in the medial or
no clear consensus on what roles these factors lateral compartment of the knee. Such altera-
play. This is likely because the mechanisms by tions in joint geometry may interfere with nutri-
which sex hormones may affect the development tion of the cartilage or alter load distribution,
and/or progression of OA are not yet under- either of which could change the biochemical
stood; there are limited hormone data available composition of the cartilage [62] .
for evaluation. Secondary analyses of data from Regardless of the exact mechanisms by which
large RCTs of HT suggest that estrogen therapy, obesity contributes to greater OA risk, a num-
but not estrogen plus progestin therapy, reduces ber of investigations have found that increased
the incidence of joint replacement in postmen- body weight contributes to the incidence of
opausal women [41,42] . It is unlikely that there OA [11,17,63–66] . Cross-sectional data from the
will be a large RCT of the effects of estrogen Framingham Osteoarthritis Study indicated
versus estrogen plus progestin therapy on joint that obese individuals (BMI  >  30.0  kg/m 2 )
health. Therefore, prospective cohort studies of are four-times more likely to have knee OA
the association between use of sex HT and OA than those with a BMI of 25.0 kg/m 2 [63] .
risk should include a focus on the type of HT. Furthermore, the RR for OA of the knee
among overweight individuals is higher in
BMI & OA women (OR: 1.8; 95% CI: 1.2–2.6) than men
Although the cause of knee OA is almost cer- (OR:1.0; 95% CI: 0.5–2.1) [11,64] . Cooper et al.
tainly multifactorial, obesity is a primary risk reported that obesity was a strong predictor of
factor [17,34,52–56] . Obesity is also the most impor- knee OA after adjusting for age and sex; obesity
tant modifiable risk factor for the development was also a significant, albeit weaker, predictor
and progression of OA [17,52,53] . Increased body of OA progression [17] .
weight is especially related to developing OA of Weight loss reduces symptoms of knee
the knee, and it has been speculated that this is OA [52] , and weight loss and exercise are recom­
because of increased mechanical loading at the mended for obese patients with knee OA by
joint [54–56] . However, the observation that obes- both the American College of Rheumatology
ity is associated with hand OA suggests mecha- and the European League Against Rheumatism
nisms other than, or in addition to, increased (EULAR) [67,68] . Observational data from the
joint loading [24,57] . One possible explanation Framingham Knee OA Study showed that a

future science group Women's Health (2010) 6(4) 605

REVIEW – Stevens-Lapsley & Kohrt

reduction in weight of 5.1  kg decreased the increase in BMI: 1.23; 95% CI: 1.05–1.45). In
risk of developing knee OA by 50% in women neutrally aligned limbs and severely malaligned
with a baseline BMI over 25.0 kg/m 2 [52] . limbs, body weight had no effect on risk of
Intervention studies support this finding, with knee OA progression in either sex. It may seem
a moderate amount of weight loss (5% of body paradoxical that increased body weight did not
weight) over the course of 18 months resulting augment the progression of OA when severe
in better self-reported physical function and malalignment was present. However, because
decreased disability, less pain and stiffness, as malalignment is a very strong risk factor for
well as faster walking speed and stair climb structural progression, it is possible that the
time compared with a weight-stable control stresses placed on cartilage with severe malalign-
group [69] . More intensive weight loss (10% ment already exceeded thresholds for cartilage
weight loss goal) resulted in similar improve- loss and were not further worsened by excessive
ments in self-reported function and walking body weight [58,74,75] .
speed [70] . A limitation of studies examining the impact
Weight loss also results in reduced compres- of obesity on OA is that many have not differ-
sive knee-joint loads, as demonstrated after an entiated between fat and fat-free mass. A small
18-month clinical trial of diet and exercise [71] . study (n = 22) of overweight adults with knee OA
In fact, the force reduction at the knee joint was evaluated whether interventions geared toward
larger than would be predicted for the weight decreasing body fat and increasing physical activ-
loss. For every 1 pound of weight loss, there was ity improved symptoms of OA [76] . Investigators
a 4‑pound reduction in knee-joint load per step. found that decreasing body fat and increasing
The theoretical accumulated reduction in knee physical activity were more important in produc-
load for a 1‑pound loss in weight would be more ing symptomatic relief of knee OA than other
than 4800 pounds per mile walked (assuming indices of obesity, such as weight loss.
1200 strides/mile). With a 10‑pound weight In summary, obesity is a primary risk factor
loss, each knee would be subjected to 48,000 for developing OA, but it is less clear how obes-
pounds less in compressive loading per mile ity influences the progression of OA. The strong
walked. A reduction of this magnitude would relationship between obesity and incident OA
likely be clinically relevant. Indeed, Felson could be secondary to increased mechanical
et al. found that a weight loss of 11.2 pounds loading at the joint, which could be further exac-
over a 10-year period was associated with a erbated by joint malalignment. Alternatively,
50% reduction in the risk of developing knee systemic and hormonal factors associated with
OA [52] . Although such observational studies obesity may adversely affect joints throughout
highlight the potential benefit of weight loss the body, and make it difficult to isolate the bio-
on joint health, long-term RCTs are needed to mechanical effects of obesity on OA. Regardless
better isolate the role of weight loss and exercise of the mechanisms for obesity-related OA, sev-
on joint protection. eral studies suggest that weight loss reduces
The Framingham study also examined symptoms of knee OA.
whether lower extremity limb alignment was
related to BMI and OA progression in women Physical activity & incidence of OA
and men [72] . Among the 394 knees examined, Physical activity and OA have a potentially
90 (22.8%) demonstrated some disease progres- complex association because a certain level of
sion. Limb alignment was strongly associated mechanical joint stress is essential for good joint
with OA progression risk; the proportion of health [77,78] , but excessive joint stress may pro-
knees showing OA progression was greater with mote the development of OA [77,79–81] . In the
poorer alignment (9.2% of neutrally aligned absence of major joint injury, there is no evi-
limbs [0–2°]; 22.3% of limbs with 3–6° of dence that regular moderate-to-vigorous physical
malalignment; 48.7% of limbs with severe activity (in amounts that are commonly recom-
malalignment [>7°]). The association of align- mended for general health benefits) increases the
ment with progression was similar between men risk of developing OA [13] . Figure 1 illustrates that
and women [72] , although in general women tend the risk for OA does not appear to increase or
to have worse lower extremity alignment than decrease with varying levels of physical activity.
men [73] . BMI was also associated with OA pro- In addition, there is limited, weak evidence from
gression, but surprisingly, the effect of BMI on observational and animal studies that low-to-
progression was limited to knees in which there moderate levels of physical activity, particularly
was moderate malalignment (OR per 2‑unit walking, may provide protection against the

606 www.futuremedicine.com future science group

 Osteoarthritis in women: effects of estrogen, obesity & physical activity – Review

1.4
Women

1.2

1.0
Risk estimate

0.8

0.6

Cheng 2000
0.4 Hootman 2003
Manninen 2001
Rogers 2002
0.2

Sedentary Low Moderate High

Physical activity category

1.4
Men

1.2

1.0
Risk estimate

0.8

0.6

Cheng 2000
0.4 Hootman 2003
Manninen 2001
Rogers 2002
0.2

Sedentary Low Moderate High

Physical activity category

Figure 1. Risk of osteoarthritis for varying levels of physical activity. Risk estimates are odds
ratios for Hootman 2003 [102] , Manninen 2001 [84] and Rogers 2002 [83] , and hazard ratios for Cheng
2000 [103] . Investigations that reported multiple, discrete categories of physical activity by sex (men vs
women) were chosen for inclusion. These representative studies support the current belief that
physical activity does not markedly increase or decrease the risk of developing osteoarthritis.

development of hip and knee OA [13] . Whether Felson et  al. examined the role of physi-
exceeding some threshold of physical activity cal activity in the development of OA in the
promotes the development of OA is unclear, but Framingham cohort (n = 1279) [82] . Participants
some level of physical activity is likely essential who did not have OA at the beginning of the
for optimal joint health [13] . study were asked on a regular basis about their

future science group Women's Health (2010) 6(4) 607

REVIEW – Stevens-Lapsley & Kohrt

recreational activities and knee pain. Over an risk for OA than those who were sedentary [82] .
average follow-up time of 9  years, new cases Therefore, while physical activity did not protect
of OA were defined three ways: radiographic against the development of OA, it also did not
OA (progression of the Kellgren and Lawrence increase the incidence of OA, even in individuals
[K/L] grade to at least 2); tibiofemoral joint who were overweight or obese.
space loss (1-grade K/L change on either lateral Another longitudinal study of women found
or anterior–posterior views); and symptomatic some evidence for a protective effect of exercise
OA (a knee that did not have the combination on incidence of OA [34] . Individuals were fol-
of symptoms and radiographic disease at base- lowed over 4 years (n = 715) for the incidence
line, but developed that combination by follow- of radiographic OA. Women who walked more
up). Results suggested that neither recreational than 5 miles per week had significantly less joint
walking, jogging, working up a sweat, nor high space narrowing (OR: 0.38; 95% CI: 0.15–0.93)
activity levels was associated with a decrease or than women who walked less than 5 miles per
increase in risk of OA compared with sedentary week (OR: 0.60; 95% CI: 0.22–1.71). Although
individuals. When compared with participants evidence is weak, other cohort and case–control
who reported no walking activity, the RR for investigations suggest that low-impact activ-
symptomatic OA was 0.96 (95% CI: 0.57–1.62) ity may have a small protective effect on the
in those who reported walking less than 6 miles ­development of OA [83,84] .
per week and 0.78 (95% CI 0.49–1.24) in those The Musculoskeletal Health subcommittee
who reported walking more than 6 miles per of the Federal Advisory Committee, who pre-
week. Similarly, neither the risk of incident pared the Physical Activity Guidelines Advisory
radiographic OA nor joint space loss was sub- Committee Report (2008), examined the scien-
stantially affected by walking for exercise. Only tific evidence from observational epidemiologic
a few subjects (n = 68) reported jogging or run- studies assessing physical activity exposure in
ning regularly for exercise, but they also did not terms of walking and OA incidence (Table 2) [13] .
appear to be at increased risk of developing knee Less information is available regarding higher
OA. Importantly, this investigation also found level activities such as running and sports par-
that people with a BMI above the median level ticipation. The report suggested that the weak
(27.7 kg/m2 for men and 25.7 kg/m2 for women) protective effect of physical activity may be
who were physically active were at no greater stronger among women than men [13] . Two

Table 2. Studies examining the association between participation in walking and risk of hip/knee osteoarthritis
from the Physical Activity Guidelines Advisory Committee report.
Study Study OA definition Walking exposure Measure of association OR (95% CI) Ref.
type
Hart et al. Cohort Incident radiographic: Walking† Joint space narrowing: No: 1.0 (referent); [34]
(1999) Joint space narrowing No: <5 miles per week Yes: 0.38 (0.15–0.93)
Osteophyte formation Yes: >5 miles per week Osteophyte formation: No: 1.0 (referent);
Yes: 0.60 (0.22–1.71)
McAlindon Cohort Radiographic knee OA Number of city blocks walked None: 1.0 (referent) [88]
et al. (1999) per day >4: 1.2 (0.4 – 3.8)
Manninen Case– Knee arthroplasty Regularly performed exercise for at Men: No: 1.0 (referent); Yes: 0.17 [84]
et al. (2001) control surgery least 2 years? (0.02–1.46)
Walking: Yes/No Women: No: 1.0 (referent); Yes: 0.32
(0.16–0.65)
Manninen Case– Knee arthroplasty Occupational walking: Low: 1.0 (referent) [101]
et al. (2002) control surgery Low, medium or high Medium: 1.0 (0.65–1.53)
High: 1.06 (0.68–1.64)
Felson et al. Cohort Radiographic, Do you walk for exercise? No: 1.0 (referent) [82]
(2007) symptomatic knee OA No <6: 0.96 (0.57–1.62)
<6 miles/week >6: 0.78 (0.49–1.24)
>6 miles/week
†
No details were provided on the question used to determine walking in Hart et al. However, another published paper from the same cohort described the walking
variable as less than versus greater than 5 miles per week.
CI: Confidence interval; OA: Osteoarthritis; OR: Odds ratio.
Adapted from the Physical Activity Guidelines Advisory Committee Report [13].

608 www.futuremedicine.com future science group

 Osteoarthritis in women: effects of estrogen, obesity & physical activity – Review

studies found that low and high levels of physi- The Physical Activity Guidelines Advisory
cal activity provided some protection from OA Committee Report (2008) concluded that
for women (not all were statistically significant there is strong evidence that endurance, resis-
due to small sample sizes), but only high levels of tance and combined endurance and resistance
activity offered protection for men [83,84] . exercise programs reduce symptomatic OA by
Participation in certain sports has been associ- decreasing pain and improving function, qual-
ated with an increased risk of OA [13] . These sports ity of life and mental health, as well as delaying
include football, soccer, track and field, basket- the onset of disability with activities of daily liv-
ball, boxing, ice hockey, orienteering, running, ing [13,91,92] . Table 3 includes exercise programs
wrestling, tennis, ballet and handball. One reason selected for the Physical Activity Guidelines
these sports are associated with increased risk for Advisory Committee Report that could be
OA may be that joint injuries are common in these replicated in community settings (e.g., group
sports and joint injury is a strong risk factor for exercise classes, home programs) with or with-
incident OA [9,11,85] . Recent studies have shown a out supervision. Most of the endurance exercise
significantly increased incidence of knee injuries training studies in Table 3 included moderate-
in female athletes when compared with their male intensity, low-impact exercise, such as walking,
counterparts [86,87] . As participation in women’s water exercise, tai chi, aerobics and cycling. Of
recreational, high school, collegiate and profes- the 15 endurance exercise studies reviewed, the
sional sports continues to increase, sex differences average exercise dose was 2.9 sessions per week
for incidence of OA may widen even further [87] . and 48 min per session for a total of 137 min
Individuals with occupations that require per week (Table 3) [13] . Most of the studies found
excessive knee bending/kneeling or involve a reduction in pain and improvements in physi-
lifting and carrying heavy loads, and who also cal function in response to exercise training, and
participate in moderate or vigorous recreational some found that exercise improved quality of
activity, may have increased risk for lower- life, muscle strength and physical activity levels.
extremity OA because of additive effects over Importantly, none of the studies found increases
time [8,88–90] . Data from the US National Health in pain, stiffness or other measures of disease
and Nutrition Examination Survey and the progression, including radiographic progression.
Framingham cohort reported that women with The Physical Activity Guidelines Advisory
jobs requiring repetitive knee bending had triple Committee Report also examined nine stud-
the rate of knee OA compared with those with ies of resistance training that involved iso-
jobs not requiring such activity (men had double tonic and/or isokinetic exercises (Table  3) [13] .
the rate), even after ­controlling for known risk Collectively, these studies demonstrated fewer
factors [8,90] . symptoms of pain, stiffness and function,
In summary, exercise at moderate levels (e.g., improved muscle strength, and provided fur-
walking) does not appear to increase the risk of ther evidence that exercise does not increase
developing OA. Instead, moderate exercise may the progression of arthritis. RCTs of com-
confer a weak protective effect, possibly favoring bined endurance and resistance exercise train-
women over men. Certain sports may increase ing found that benefits were similar to those
the risk of OA, possibly because of greater like- reported for endurance-only and resistance-
lihood for joint injury. Similarly, occupations only programs [13] . Thus, there is strong and
with high physical demands may predispose consistent evidence from these relatively short-
individuals to OA. Finally, individuals with term exercise intervention trials (up to 2 years)
higher BMIs can safely participate in moderate- that exercise improves pain, function, muscle
intensity physical a­ctivity without increased risk strength and mental health with no increase in
for knee OA. disease severity.
Based on this evidence, the Physical Activity
Physical activity & progression of OA Guidelines Advisory Committee Report con-
Considerable attention has focused on how cluded that people with OA can experience
physical activity affects disease progression improvements in pain and physical function
once OA is present. Exercise intervention trials by engaging in moderate-to-vigorous intensity,
have evaluated whether endurance and/or resis- low-impact physical activities for approximately
tance exercise training slows the progression of 150 min per week [13] . Walking was recommended
OA (i.e.,  radiographic changes) and whether as particularly appropriate. The consensus state-
these exercise strategies improve or exacerbate ment from the International Osteoarthritis
OA symptoms and quality of life [13,91–94] . Research Society (OARSI) is in agreement with

future science group Women's Health (2010) 6(4) 609

610
Table 3. Summary of descriptive characteristics of the randomized controlled trials of exercise among persons with arthritis or other rheumatic
conditions from the Physical Activity Guidelines Advisory Committee report.
Study type Studies Average (mean) characteristics of interventions Significant findings
(n) (number of studies/outcome)
Intervention Control Duration of Frequency Duration per Total prescribed
subjects (n) subjects (n) intervention per week session (min) dose (min/week)
[range] [range] (weeks) [range] [range] [range] [range]
Endurance 17† 50 [17–144] 45 [16–149] 23.9 [8–72] 2.9 [2–5] 47.8 [20–60] 137 [60–180] 10 ↓ pain
vs control 8 ↑ function
1 ↑ quality of life
4 ↑ self-efficacy
4 ↑ muscle strength
2 ↑ physical activity
3 ↓ symptoms (other than pain)
REVIEW – Stevens-Lapsley & Kohrt

4 ↑ mental/emotional health
5 ↑ or no change in symptoms/disease activity
Resistance vs 9 54 [10–146] 55 [10–149] 50.9 [8–96] 2.6 [2–3] 52.5 [30–60] 145 [60–180] 5 ↓ pain
control 5 ↑ function
6 ↑ muscle strength
3 ↓ stiffness
3 ↓ disease activity
4 ↓ disability
1 ↑ ROM
Combination 5 62 [25–151] 64 [25–158] 44.0 [12–104] 3.0 [2–5] 55.0 [30–75] 156 [120–180] 1 ↓ pain
vs control 2 ↑ function

www.futuremedicine.com
2 ↑ muscle strength
2 ↑ fitness/perceived exertion
2 ↑ no change in disease activity
1 ↑ mental health
1 ↓ body weight
All studies 24‡ 54 52 39.6 2.8 51.8 146
All studies implemented exercise interventions of at least moderate intensity.
†
The endurance group had 15 individual studies, but 17 actual exercise versus control comparisons.
‡
Review included 24 individual studies, two studies compared multiple exercise groups versus a non-exercise control group and may be counted separately under the rows for the endurance, resistance and
combination studies.
Adapted from the Physical Activity Guidelines Advisory Committee Report [11].

future science group

 Osteoarthritis in women: effects of estrogen, obesity & physical activity – Review

these guidelines [95] . Of the 11 nonpharmaco- Future perspective

logical OARSI recommendations, there was a This review highlighted the potential contribu-
strong recommendation that all patients with tions of estrogen, obesity and physical activity on
hip and/or knee OA should engage in endurance the incidence and progression of OA, with empha-
and resistance training and range-of-motion sis on sex-specificity when possible. Despite the
exercises. OARSI reported a moderate effect higher incidence of OA in women than men, few
of exercise on pain relief (pooled effect sizes of studies have adequately examined sex-specific dif-
0.52; 95% CI: 0.34–0.70 for endurance exercise ferences. Why women are more susceptible to OA
vs 0.32; 95% CI: 0.23–0.42 for resistance exer- remains unclear and necessitates further research.
cise) [95] . The American Geriatrics Association Although many studies have explored the relation-
Consensus Practice Statement [96] , OASIS state- ships of estrogen and reproductive history with the
ment [97] , MOVE Consensus [98] , EULAR rec- risk of OA or joint replacement, there is no clear
ommendations [67,99,100] , and American College consensus on the role of these factors. Obesity
of Rheumatology guidelines [67] are also in increases the risk for developing OA, but whether
agreement with the Physical Activity Guidelines obesity affects women and men differently has not
Advisory Committee Report. been thoroughly investigated. Finally, moderate
In summary, there is no indication that low- to levels of physical activity do not appear to increase
moderate-intensity exercise causes increased pain the risk of developing OA, even in overweight or
or disability in men or women with OA. In fact, obese adults, and may even have a weak protective
there is strong evidence that endurance, resistance effect. Once OA has developed, moderate levels
and combined endurance/resistance exercise pro- of physical activity may mitigate some symptoms
grams reduce symptomatic OA by decreasing pain associated with OA progression in both men and
and improving function, quality of life and men- women. Future investigations should focus on
tal health, as well as delaying the onset of dis- sex-specific mechanisms for the development of
ability with activities of daily living. Longer dura- OA and determine whether there are modifiable
tion exercise programs (>2 years) have not been factors that can be targeted through prevention
evaluated, so conclusions regarding effectiveness and treatment strategies to mitigate the increased
of longer intervention are not currently available. prevalence of OA in women.

Executive summary
Osteoarthritis
• Arthritis is the leading cause of physical disability in the USA.
• The most common form of arthritis is osteoarthritis (OA), which affects nearly 27 million people in the USA.
• Women have a higher prevalence of OA than men, especially in the knee.
• Clinical signs of symptomatic OA include slowly developing joint pain, stiffness and enlargement with limitation of motion.
• Radiographic signs of OA include osteophytes, joint space narrowing and sclerosis.
• Risk factors for the development of OA include: obesity, female sex, older age, previous knee trauma and occupational load.
• Incidence and progression of symptomatic and radiographic knee OA may involve different processes.
Sex-specificity of OA & hormonal influences
Estrogen-based hormone therapy
• The incidence of OA rises more steeply in women than in men after 50 years of age, suggesting an association with the menopausal
withdrawal of ovarian hormones. Evidence to support this is mixed.
• Some studies, but not all, have found a lower incidence or prevalence of knee and hip OA in hormone therapy (HT) users, suggesting
that estrogens may have a protective effect on joints.
• Few RCTs have been performed to evaluate the effects of HT on OA, and none has evaluated HT on OA risk as a primary outcome. The
available evidence suggests that estrogen, but not estrogen plus progestin, therapy reduces OA risk.
Previous hormonal contraceptive use
• There appears to be no significant association of the use or duration of use of hormonal contraceptives with OA.
Age of menopause & age at menarche
• The increased incidence of OA around the time of menopause suggests that the development of OA is linked with the menopausal
withdrawal of ovarian hormones.
• Attempts to establish a temporal relationship between the onset of OA and menopause have been inconclusive.
• The bulk of evidence suggests that age at menopause and menopausal status (pre/peri versus post) are not related to incidence of OA.
• Only a few studies have investigated the relationship between age of menarche and OA risk; some have reported that a younger age at
menarche may be related to increased incidence of OA, while others have not found such an association.
Parity
• Few studies have found an association between parity and risk for OA.

future science group Women's Health (2010) 6(4) 611

REVIEW – Stevens-Lapsley & Kohrt

Executive summary (cont.)

BMI & OA
• Obesity is a primary risk factor for developing OA.
• Mechanical loading of joints and/or systemic and hormonal factors responsible for bone and cartilage metabolism may explain why
people with higher BMIs are at greater risk for developing OA.
• Weight loss reduces symptoms of knee OA, possibly because of reduced compressive knee-joint loads.
Physical activity & incidence of OA
• Physical activity and OA have a potentially complex association because a certain level of mechanical joint stress is essential for good
joint health, but excessive joint stress may promote the development of OA.
• There is no evidence that regular moderate to vigorous physical activity, in amounts that are commonly recommended for general
health benefits, increases the risk of developing OA.
• Moderate physical activity may have a weak protective benefit for developing OA.
• Whether exceeding some threshold of physical activity promotes the development of OA is unclear.
• Physical activity does not appear to increase the incidence of OA in individuals who are overweight.
Physical activity & progression of OA
• There is strong evidence that endurance, resistance and combined endurance and resistance exercise programs reduce symptomatic OA
by 1) decreasing pain, 2) improving function, quality of life and mental health, and 3) delaying the onset of disability with activities of
daily living.
• The Physical Activity Guidelines Advisory Committee Report concluded that people with OA can experience improvements in pain and
physical function by engaging in moderate-to-vigorous intensity, low-impact physical activities for approximately 150 min per week.
Walking is recommended as a particularly appropriate exercise.
Future perspective
• Despite the higher incidence of OA in women than men, few studies have adequately examined sex-specific differences. The reasons
why women are more susceptible to OA remain unclear and necessitate further research. Therefore, future investigation should
focus on sex-specific mechanisms for the development of OA and determine whether there are modifiable factors (e.g., estrogen,
obesity, physical activity) that can be targeted through prevention and treatment strategies to mitigate the increased incidence and
prevalence of OA in women.
• Prospective RCTs are needed to investigate the association of menopausal estrogen-based HT with OA. To date, few RCTs have
evaluated the effects of HT on OA from an intervention perspective, and none have evaluated OA risk as a primary outcome.
• Further investigation of how obesity impacts OA via either systemic hormonal influences or the biomechanical effects of altered joint
loading is also an area for further investigation, especially with respect to how these factors may affect women more than men.
• Prospective investigations evaluating how physical activity levels contribute to OA incidence and progression in men versus women
are warranted. More specifically, a more comprehensive understanding of how different physical activity levels contribute to OA
incidence would provide valuable information for clinical guidelines to prevent OA. Similarly, investigations of how physical activity
intervention (especially longer than 2 years) affects OA progression would provide alternatives to pharmacological options for
treating OA.
• Finally, a better understanding of how the complex interactions between obesity, weight loss and energy expenditure through
physical activity in men and women contribute to OA would enhance current treatment guidelines.

4. Hunter DJ: Imaging insights on the

Financial & competing Bibliography epidemiology and pathophysiology of
interests disclosure Papers of special note have been highlighted as:
osteoarthritis. Rheum. Dis. Clin. North Am. 35,
• of interest
The authors have no relevant affiliations or finan- 447–463 (2009).
•• of considerable interest
cial involvement with any organization or entity 5. Ai F, Yu C, Zhang W, Morelli JN,
1. Lawrence RC, Felson DT, Helmick CG et al.:
with a financial interest in or financial conflict Kacher D, Li X: MR imaging of knee
Estimates of the prevalence of arthritis and
with the subject matter or materials discussed in osteoarthritis and correlation of findings with
other rheumatic conditions in the United
the manuscript. This includes employment, consul- reported patient pain. J. Huazhong Univ. Sci.
States. Part II. Arthritis Rheum. 58, 26–35
tancies, honoraria, stock ownership or options, Technolog. Med. Sci. 30(2), 248–524 (2010).
(2008).
expert testimony, grants or patents received or 6. Hannan MT, Anderson JJ, Pincus T,
2. Hutton CW: Osteoarthritis: the cause not
p­ending, or royalties. Felson DT: Educational attainment and
result of joint failure? Ann. Rheum. Dis.
osteoarthritis: differential associations
No writing assistance was utilized in the produc- 48, 958–961 (1989).
with radiographic changes and symptom
tion of this manuscript. 3. Arden N, Nevitt MC: Osteoarthritis: reporting. J. Clin. Epidemiol. 45, 139–147
epidemiology. Best Pract. Res. Clin. (1992).
Rheumatol. 20, 3–25 (2006).
7. Felson DT: An update on the pathogenesis
• Nice review of the epidemiology   and epidemiology of osteoarthritis. Radiol.
of osteoarthritis. Clin. North Am. 42, 1–9 (2004).

612 www.futuremedicine.com future science group

 Osteoarthritis in women: effects of estrogen, obesity & physical activity – Review

8. Anderson JJ, Felson DT: Factors associated 20. Dillon CF, Rasch EK, Gu Q, Hirsch R: 32. Reginster JY, Kvasz A, Bruyere O, Henrotin Y:
with osteoarthritis of the knee in the first Prevalence of knee osteoarthritis in the United Is there any rationale for prescribing hormone
national Health and Nutrition Examination States: arthritis data from the Third National replacement therapy (HRT) to prevent or to
Survey (HANES I). Evidence for an Health and Nutrition Examination Survey treat osteoarthritis? Osteoarthritis Cartilage 11,
association with overweight, race, and 1991–94. J. Rheumatol. 33, 2271–2279 (2006). 87–91 (2003).
physical demands of work. Am. J. Epidemiol. 21. Felson DT, Zhang Y, Hannan MT et al.: 33. Hannan MT, Felson DT, Anderson JJ,
128, 179–189 (1988). The incidence and natural history of knee Naimark A, Kannel WB: Estrogen use and
9. Felson DT: Risk factors for osteoarthritis: osteoarthritis in the elderly. The Framingham radiographic osteoarthritis of the knee in
understanding joint vulnerability. Osteoarthritis Study. Arthritis Rheum. women. The Framingham Osteoarthritis
Clin. Orthop. Relat. Res. S16–21 (2004). 38, 1500–1505 (1995). Study. Arthritis Rheum. 33, 525–532 (1990).
10. Felson DT: Developments in the clinical 22. Brooks PM: The burden of musculoskeletal • Results of osteoarthritis incidence in female
understanding of osteoarthritis. Arthritis Res. disease--a global perspective. Clin. Rheumatol. participants of the Framingham
Ther. 11, 203 (2009). 25, 778–781 (2006). Osteoarthritis Study evaluated with
11. Felson DT, Lawrence RC, Dieppe PA et al.: 23. Cho HJ, Chang CB, Yoo JH, Kim SJ, weight-bearing radiographs of the knee.
Osteoarthritis: new insights. Part 1: the Kim TK: Gender differences in the correlation 34. Hart DJ, Doyle DV, Spector TD: Incidence and
disease and its risk factors. Ann. Intern. Med. between symptom and radiographic severity in risk factors for radiographic knee osteoarthritis
133, 635–646. (2000). patients with knee osteoarthritis. Clin. Orthop. in middle-aged women: the Chingford Study.
12. Blagojevic M, Jinks C, Jeffery A, Jordan KP: Relat. Res. DOI:10.1007/s11999-010-1282-z Arthritis Rheum. 42, 17–24 (1999).
Risk factors for onset of osteoarthritis of the (2010) (Epub ahead of print).
• Chingford Study results examining the
knee in older adults: a systematic review and 24. Hannan MT: Epidemiologic perspectives on
natural history, role of risk factors and
meta-analysis. Osteoarthritis Cartilage 18(1), women and arthritis: an overview. Arthritis
incidence of knee osteoarthritis in a
24–33 (2010). Care Res. 9, 424–434 (1996).
prospective study of women from a
• Recent systematic review of the risk factors 25. Kapstad H, Rustoen T, Hanestad BR et al.: population cohort.
for the incidence of osteoarthritis. Changes in pain, stiffness and physical
35. Nevitt MC, Felson DT: Sex hormones and the
function in patients with osteoarthritis waiting
13. Physical Activity Guidelines Advisory risk of osteoarthritis in women:
for hip or knee joint replacement surgery.
Committee. Physical Activity Guidelines epidemiological evidence. Ann. Rheum. Dis.
Osteoarthritis Cartilage 15, 837–843 (2007).
Advisory Committee Report, 2008. US 55, 673–676 (1996).
Department of Health and Human Services, 26. Kelly KD, Voaklander DC, Johnston DWC,
36. Samanta A, Jones A, Regan M, Wilson S,
Washington, DC, USA (2008). Newman SC, Suarez-Almazor ME: Change in
Doherty M: Is osteoarthritis in women affected
pain and function while waiting for major joint
•• Excellent overview the impact of physical by hormonal changes or smoking?
arthroplasty. J. Arthroplasty 16, 351–359 (2001).
activity on the incidence and progression   Br. J. Rheumatol. 32, 366–370 (1993).
of osteoarthritis. 27. Srikanth VK, Fryer JL, Zhai G et al.: A
37. Spector TD, Nandra D, Hart DJ, Doyle DV:
meta-analysis of sex differences prevalence,
14. de Klerk BM, Schiphof D, Groeneveld FP Is hormone replacement therapy protective for
incidence and severity of osteoarthritis.
et al.: No clear association between female hand and knee osteoarthritis in women?
Osteoarthritis Cartilage 13, 769–781 (2005).
hormonal aspects and osteoarthritis of the The Chingford Study. Ann. Rheum. Dis. 56,
hand, hip and knee: a systematic review. 28. de Klerk BM, Schiphof D, Groeneveld FP 432–434 (1997).
Rheumatology (Oxford) 48, 1160–1165 (2009). et al.: Limited evidence for a protective
38. Zhang Y, McAlindon TE, Hannan MT et al.:
effect of unopposed oestrogen therapy for
•• Recent systematic review investigating the Estrogen replacement therapy and worsening
osteoarthritis of the hip: a systematic review.
assumed association between osteoarthritis of radiographic knee osteoarthritis: the
Rheumatology (Oxford) 48, 104–112 (2009).
and hormonal influences in women. Framingham Study. Arthritis Rheum. 41,
29. Liu B, Balkwill A, Cooper C et al.: 1867–1873 (1998).
15. Sharma L, Lou C, Felson DT et al.: Laxity in Reproductive history, hormonal factors and the
healthy and osteoarthritic knees. Arthritis • First prospective cohort study
incidence of hip and knee replacement for
Rheum. 42, 861–870 (1999). osteoarthritis in middle-aged women. (Framingham) to examine the effects of
16. Sharma L, Dunlop DD, Cahue S, Song J, Ann. Rheum. Dis. 68, 1165–1170 (2009). estrogen-replacement therapy on
Hayes KW: Quadriceps strength and radiographic knee osteoarthritis.
•• One of the largest investigations of
osteoarthritis progression in malaligned and reproductive history and use of hormonal 39. Parazzini F: Menopausal status, hormone
lax knees. Ann. Intern. Med. 138, 613–619 replacement therapy use and risk of self-
therapies on the risk of hip and knee joint
(2003). reported physician-diagnosed osteoarthritis in
replacement for osteoarthritis.
17. Cooper C, Snow S, McAlindon TE et al.: Risk women attending menopause clinics in Italy.
30. Richmond RS, Carlson CS, Register TC, Maturitas 46, 207–212 (2003).
factors for the incidence and progression of
Shanker G, Loeser RF: Functional estrogen
radiographic knee osteoarthritis. Arthritis 40. Karlson EW, Mandl LA, Aweh GN et al.:
receptors in adult articular cartilage: estrogen
Rheum. 43, 995–1000 (2000). Total hip replacement due to osteoarthritis: the
replacement therapy increases chondrocyte
18. Doherty M: Risk factors for progression of knee importance of age, obesity, and other
synthesis of proteoglycans and insulin-like
osteoarthritis. Lancet 358, 775–776 (2001). modifiable risk factors. Am. J. Med. 114, 93–98
growth factor binding protein 2. Arthritis
(2003).
19. Belo JN, Berger MY, Reijman M, Koes BW, Rheum. 43, 2081–2090 (2000).
Bierma-Zeinstra SM: Prognostic factors of 41. Cirillo DJ, Wallace RB, Wu L, Yood RA: Effect
31. Roman-Blas JA, Castaneda S, Largo R,
progression of osteoarthritis of the knee: a of hormone therapy on risk of hip and knee joint
Herrero-Beaumont G: Osteoarthritis
systematic review of observational studies. replacement in the Women’s Health Initiative.
associated with estrogen deficiency. Arthritis
Arthritis Rheum. 57, 13–26 (2007). Arthritis Rheum. 54, 3194–3204 (2006).
Res. Ther. 11, 241 (2009).

future science group Women's Health (2010) 6(4) 613

REVIEW – Stevens-Lapsley & Kohrt

•• Results from the Women’s Health Initiative • Framingham cohort study describing a 65. Felson DT, Anderson JJ, Naimark A,
on the effect of hormone therapy on risk of positive relationship between weight loss Walker AM, Meenan RF: Obesity and knee
hip and knee joint replacement. and decreased risk for symptomatic knee osteoarthritis. The Framingham Study. Ann.
osteoarthritis in women. Intern. Med. 109, 18–24 (1988).
42. Nevitt MC, Felson DT, Williams EN, Grady D:
The effect of estrogen plus progestin on knee 53. Felson DT: The epidemiology of knee 66. Spector TD, Hart DJ, Doyle DV: Incidence
symptoms and related disability in osteoarthritis: results from the Framingham and progression of osteoarthritis in women
postmenopausal women: The Heart and Osteoarthritis Study. Semin. Arthritis Rheum. with unilateral knee disease in the general
Estrogen/Progestin Replacement Study, a 20, 42–50 (1990). population: the effect of obesity. Ann. Rheum.
randomized, double-blind, placebo-controlled Dis. 53, 565–568 (1994).
• Early results of the Framingham  
trial. Arthritis Rheum. 44, 811–818 (2001). 67. Recommendations for the medical
Knee Osteoarthritis study examining the
•• Randomized trial data investigating the prevalence of radiographic and management of osteoarthritis of the hip and
association of hormone treatment with knee knee: 2000 update. American College of
symptomatic knee osteoarthritis in
or hip osteoarthritis as part of the Heart and Rheumatology Subcommittee on
independently living elderly.
Estrogen/Progestin Replacement Study. Osteoarthritis Guidelines. Arthritis Rheum. 43,
54. Gillespie GN, Porteous AJ: Obesity and knee 1905–1915 (2000).
43. Tepper S, Hochberg MC: Factors associated arthroplasty. Knee 14, 81–86 (2007).
with hip osteoarthritis: data from the First • American College of Rheumatology
55. Sturmer T, Gunther KP, Brenner H: Obesity,
National Health and Nutrition Examination recommendations for hip and knee
overweight and patterns of osteoarthritis: the
Survey (NHANES-I). Am. J. Epidemiol. 137, osteoarthritis management.
Ulm Osteoarthritis Study. J. Clin. Epidemiol.
1081–1088 (1993). 68. Pendleton A, Arden N, Dougados M et al.:
53, 307–313 (2000).
44. Spector TD, Campion GD: Generalised EULAR recommendations for the
56. Bourne R, Mukhi S, Zhu N, Keresteci M,
osteoarthritis: a hormonally mediated disease. management of knee osteoarthritis: report of a
Marin M: Role of obesity on the risk for total
Ann. Rheum. Dis. 48, 523–527 (1989). task force of the Standing Committee for
hip or knee arthroplasty. Clin. Orthop. Relat.
International Clinical Studies Including
45. Liu B, Balkwill A, Banks E et al.: Relationship Res. 465, 185–188 (2007).
Therapeutic Trials (ESCISIT). Ann. Rheum.
of height, weight and body mass index to the
57. Sowers MF, Hochberg M, Crabbe JP Dis. 59, 936–944 (2000).
risk of hip and knee replacements in
et al.: Association of bone mineral density
middle-aged women. Rheumatology (Oxford) 69. Miller GD, Rejeski WJ, Williamson JD
and sex hormone levels with osteoarthritis
46, 861–867 (2007). et al.: The Arthritis, Diet and Activity
of the hand and knee in premenopausal
Promotion Trial (ADAPT): design,
46. Kalichman L, Kobyliansky E: Age, body women. Am. J. Epidemiol. 143, 38–47
rationale, and baseline results. Control Clin.
composition, and reproductive indices as (1996).
Trials 24, 462–480 (2003).
predictors of radiographic hand osteoarthritis
58. Miyazaki T, Wada M, Kawahara H et al.:
in Chuvashian women. Scand J. Rheumatol. 70. Miller GD, Nicklas BJ, Davis C et al.:
Dynamic load at baseline can predict
36, 53–57 (2007). Intensive weight loss program improves
radiographic disease progression in medial
physical function in older obese adults with
47. Cooley HM, Stankovich J, Jones G: compartment knee osteoarthritis. Ann. Rheum.
knee osteoarthritis. Obesity (Silver Spring) 14,
The association between hormonal and Dis. 61, 617–622 (2002).
1219–1230 (2006).
reproductive factors and hand osteoarthritis.
59. Dumond H, Presle N, Terlain B et al.: Evidence
Maturitas 45, 257–265 (2003). 71. Messier SP, Gutekunst DJ, Davis C, DeVita P:
for a key role of leptin in osteoarthritis. Arthritis
Weight loss reduces knee-joint loads in
48. Remsberg KE, Demerath EW, Schubert CM Rheum. 48, 3118–3129 (2003).
overweight and obese older adults with knee
et al.: Early menarche and the development of
60. Abramson SB, Attur M: Developments in the osteoarthritis. Arthritis Rheum. 52, 2026–2032
cardiovascular disease risk factors in adolescent
scientific understanding of osteoarthritis. (2005).
girls: the Fels Longitudinal Study. J. Clin.
Arthritis Res. Ther. 11, 227 (2009).
Endocrinol. Metab. 90, 2718–2724 (2005). • The Arthritis, Diet, and Activity Promotion
61. Kohrt WM, Malley MT, Dalsky GP, Trial investigating whether long-term
49. Dawson J, Juszczak E, Thorogood M et al.: An
Holloszy JO: Body composition of healthy exercise (18 months) and dietary weight
investigation of risk factors for symptomatic
sedentary and trained, young and older men loss are more effective than usual care in
osteoarthritis of the knee in women using a life
and women. Med. Sci. Sports Exerc. 24,
course approach. J. Epidemiol. Community improving physical function, pain and
832–837 (1992).
Health 57, 823–830 (2003). mobility in older overweight and obese
62. Bullough PG: The geometry of diarthrodial adults with knee osteoarthritis.
50. Dennison EM, Arden NK, Kellingray S et al.:
joints, its physiologic maintenance, and the
Hormone replacement therapy, other 72. Felson DT, Goggins J, Niu J, Zhang Y,
possible significance of age-related changes in
reproductive variables and symptomatic hip Hunter DJ: The effect of body weight on
geometry-to-load distribution and the
osteoarthritis in elderly white women: a case– progression of knee osteoarthritis is dependent
development of osteoarthritis. Clin. Orthop.
control study. Br. J. Rheumatol. 37, 1198–1202 on alignment. Arthritis Rheum. 50, 3904–3909
Relat. Res.61–66 (1981).
(1998). (2004).
63. Felson DT, Naimark A, Anderson J et al.:
51. Spector TD, Roman E, Silman AJ: The pill, • Evaluated whether the effect of body weight
The prevalence of knee osteoarthritis in
parity, and rheumatoid arthritis. Arthritis on progression of knee osteoarthritis
the elderly. The Framingham Osteoarthritis
Rheum. 33, 782–789 (1990). differs depending on the degree of  
Study. Arthritis Rheum. 30, 914–918 (1987).
52. Felson DT, Zhang Y, Anthony JM, limb malalignment.
64. Felson DT, Zhang Y, Hannan MT et al.:
Naimark A, Anderson JJ: Weight loss reduces 73. Nguyen AD, Shultz SJ: Sex differences in
Risk factors for incident radiographic knee
the risk for symptomatic knee osteoarthritis in clinical measures of lower extremity alignment.
osteoarthritis in the elderly: the Framingham
women. The Framingham Study. Ann. Intern. J. Orthop. Sports Phys. Ther. 37, 389–398
Study. Arthritis Rheum. 40, 728–733 (1997).
Med. 116, 535–539 (1992). (2007).

614 www.futuremedicine.com future science group

 Osteoarthritis in women: effects of estrogen, obesity & physical activity – Review

74. Sharma L, Song J, Felson DT et al.: The role of 85. Felson DT, Zhang Y: An update on the 96. Exercise prescription for older adults with
knee alignment in disease progression and epidemiology of knee and hip osteoarthritis osteoarthritis pain: consensus practice
functional decline in knee osteoarthritis. with a view to prevention. Arthritis Rheum. 41, recommendations. A supplement to the AGS
JAMA 286, 188–195 (2001). 1343–1355 (1998). Clinical Practice Guidelines on the
75. Felson DT, McLaughlin S, Goggins J et al.: 86. Gwinn DE, Wilckens JH, McDevitt ER, management of chronic pain in older adults.
Bone marrow edema and its relation to Ross G, Kao TC: The relative incidence of J. Am. Geriatr. Soc. 49, 808–823 (2001).
progression of knee osteoarthritis. Ann. Intern. anterior cruciate ligament injury in men and 97. Vignon E, Valat JP, Rossignol M et al.:
Med. 139, 330–336 (2003). women at the United States Naval Academy. Osteoarthritis of the knee and hip and activity:
76. Toda Y, Toda T, Takemura S et al.: Change in Am. J. Sports Med. 28, 98–102 (2000). a systematic international review and synthesis
body fat, but not body weight or metabolic 87. Hutchinson MR, Ireland ML: Knee injuries in (OASIS). Joint Bone Spine 73, 442–455
correlates of obesity, is related to symptomatic female athletes. Sports Med. 19, 288–302 (1995). (2006).
relief of obese patients with knee osteoarthritis 88. McAlindon TE, Wilson PW, Aliabadi P, • Systematic review to evaluate which
after a weigh tcontrol program. J. Rheumatol. Weissman B, Felson DT: Level of physical activities in four domains, daily life,
25, 2181–2186 (1998). activity and the risk of radiographic and exercises, sports and occupational
77. Otterness IG, Eskra JD, Bliven ML et al.: symptomatic knee osteoarthritis in the elderly: activities, should be recommended, in favor
Exercise protects against articular cartilage the Framingham study. Am. J. Med. 106, or against, for the patient suffering from
degeneration in the hamster. Arthritis Rheum. 151–157 (1999). knee or hip osteoarthritis
41, 2068–2076 (1998). 89. Vingard E, Alfredsson L, Malchau H: 98. Roddy E, Zhang W, Doherty M et al.:
78. Galois L, Etienne S, Grossin L et al.: Osteoarthrosis of the hip in women and its Evidence-based recommendations for the role
Moderate-impact exercise is associated with relationship to physical load from sports of exercise in the management of osteoarthritis
decreased severity of experimental activities. Am. J. Sports Med. 26, 78–82 (1998). of the hip or knee – the MOVE consensus.
osteoarthritis in rats. Rheumatology (Oxford) 90. Felson DT, Hannan MT, Naimark A et al.: Rheumatology (Oxford) 44, 67–73 (2005).
42, 692–693; author reply 693–694 (2003). Occupational physical demands, knee bending, 99. Jordan KM, Arden NK, Doherty M
79. Lane NE, Buckwalter JA: Exercise and and knee osteoarthritis: results from the et al.: EULAR Recommendations 2003:
osteoarthritis. Curr. Opin. Rheumatol. 11, Framingham Study. J. Rheumatol. 18, an evidence based approach to the
413–416 (1999). 1587–1592 (1991). management of knee osteoarthritis:
80. Spector TD, Harris PA, Hart DJ et al.: 91. Ettinger WH Jr, Burns R, Messier SP et al.: report of a Task Force of the Standing
Risk of osteoarthritis associated with long-term A randomized trial comparing aerobic exercise Committee for International Clinical
weight-bearing sports: a radiologic survey of and resistance exercise with a health education Studies Including Therapeutic Trials
the hips and knees in female ex-athletes and program in older adults with knee osteoarthritis. (ESCISIT). Ann. Rheum. Dis. 62, 1145–1155
population controls. Arthritis Rheum. 39, The Fitness Arthritis and Seniors Trial (FAST). (2003).
988–995 (1996). JAMA 277, 25–31 (1997). 100. Zhang W, Doherty M, Arden N et al.:
81. Kiviranta I, Tammi M, Jurvelin J et al.: • Investigation exploring the effects of EULAR evidence based recommendations for
Articular cartilage thickness and structured exercise programs on self- the management of hip osteoarthritis: report of
glycosaminoglycan distribution in the young a task force of the EULAR Standing
reported disability in older adults with
canine knee joint after remobilization of the Committee for International Clinical Studies
knee osteoarthritis.
immobilized limb. J. Orthop. Res. 12, 161–167 Including Therapeutics (ESCISIT). Ann.
92. Lange AK, Vanwanseele B, Fiatarone Singh MA: Rheum. Dis. 64, 669–681 (2005).
(1994).
Strength training for treatment of
82. Felson DT, Niu J, Clancy M et al.: • European League against Rheumatism
osteoarthritis of the knee: a systematic review.
Effect of recreational physical activities Arthritis Rheum. 59, 1488–1494 (2008). recommendations for management of 
on the development of knee osteoarthritis in knee osteoarthritis.
• Systematic review to assess the effectiveness
older adults of different weights: the 101. Manninen P, Heliovaara M, Riihimaki H,
Framingham Study. Arthritis Rheum. 57, 6–12 of isolated resistance training on arthritis
Suoma-Iainen O: Physical workload
(2007). symptoms, physical performance and
and the risk of severe knee osteoarthritis.
psychological function in people with  
•• Investigation within the Framingham study Scand. J. Work Environ. Health 28, 25–32
knee osteoarthritis.
of whether BMI or physical activity (2002).
93. Urquhart DM, Soufan C, Teichtahl AJ et al.:
increase the risk of osteoarthritis. 102. Hootman JM, Macera CA, Helmick CG,
Factors that may mediate the relationship
83. Rogers LQ, Macera CA, Hootman JM, Blair SN: Influence of physical activity-related
between physical activity and the risk for
Ainsworth BE, Blairi SN: The association joint stress on the risk of self-reported hip/
developing knee osteoarthritis. Arthritis Res.
between joint stress from physical activity and knee osteoarthritis: a new method to quantify
Ther. 10, 203 (2008).
self-reported osteoarthritis: an analysis of the physical activity. Prev. Med. 36, 636–644
94. Bosomworth NJ: Exercise and knee (2003).
Cooper Clinic data. Osteoarthritis Cartilage
10, 617–622 (2002). osteoarthritis: benefit or hazard? Can. Fam.
103. Cheng Y, Macera CA, Davis DR et al.:
Physician 55, 871–878 (2009).
84. Manninen P, Riihimaki H, Heliovaara M, Physical activity and self-reported, physician-
95. Zhang W, Moskowitz RW, Nuki G et al.: diagnosed osteoarthritis: is physical activity a
Suomalainen O: Physical exercise and risk of
severe knee osteoarthritis requiring OARSI recommendations for the management risk factor? J. Clin. Epidemiol. 53, 315–322
arthroplasty. Rheumatology (Oxford) 40, of hip and knee osteoarthritis, Part II: OARSI (2000).
432–437 (2001). evidence-based, expert consensus guidelines.
Osteoarthritis Cartilage 16, 137–162 (2008).

future science group Women's Health (2010) 6(4) 615