Obesity and Other Modifiable Factors for
Physical Inactivity Measured by Accelerometer in
Adults With Knee Osteoarthritis
JUNGWHA LEE,1JING SONG,1JENNIFER M. HOOTMAN,2PAMELA A. SEMANIK,1
ROWLAND W. CHANG,1LEENA SHARMA,1LINDA VAN HORN,1JOAN M. BATHON,3
CHARLES B. EATON,4MARC C. HOCHBERG,5REBECCA JACKSON,6C. KENT KWOH,7
W. JERRY MYSIW,6MICHAEL NEVITT,8AND DOROTHY D. DUNLOP1
Objective. To investigate the public health impact of obesity and other modifiable risk factors related to physical
inactivity in adults with knee osteoarthritis (OA).
Methods. The frequency of inactivity as defined by the US Department of Health and Human Services was assessed from
objective accelerometer monitoring of 1,089 participants (ages 49–84 years) with radiographic knee OA during the
Osteoarthritis Initiative 48-month visit (2008–2010). The relationship between modifiable factors (weight status, dietary
fat, fiber, smoking, depressive symptoms, knee function, knee pain, and knee confidence) with inactivity was assessed
using odds ratios (ORs) and attributable fractions (AFs), controlling for descriptive factors (age, sex, race, education, lives
alone, employment, frequent knee symptoms, and comorbidity).
Results. Almost half (48.9%) of participants with knee OA were inactive. Being overweight (OR 1.8, 95% confidence
interval [95% CI] 1.2–2.5) or obese (OR 3.9, 95% CI 2.6–5.7), having inadequate dietary fiber intake (OR 1.6, 95% CI
1.2–2.2), severe knee dysfunction (OR 1.9, 95% CI 1.3–2.8), and severe pain (OR 1.7, 95% CI 1.1–2.5) were significantly
related to inactivity, controlling for descriptive factors. Modifiable factors with significant average AFs were being
overweight or obese (AF 23.8%, 95% CI 10.5–38.6%) and inadequate dietary fiber (AF 12.1%, 95% CI 0.1–24.5%),
controlling for all factors.
Conclusion. Being obese or overweight, the quality of the diet, severe pain, and severe dysfunction are significantly
associated with physical inactivity in adults with knee OA. All components should be considered in designing physical
activity interventions that target arthritis populations with low activity levels.
Arthritis is a growing public health problem in the US,
affecting almost 50 million adults (1). The prevalence of
activity limitations attributable to arthritis is more than 21
million. Osteoarthritis (OA) affecting the knee is currently
a leading cause of disability in adults (2–4). It is widely
recognized that physical activity offers important benefits
to persons with arthritis, including knee OA. Physical
ClinicalTrials.gov identifier: NCT0080171.
The findings and conclusions herein are those of the au-
thors and do not necessarily represent the official position
of the Centers for Disease Control and Prevention or the
Supported in part by the National Institute for Arthritis
and Musculoskeletal Diseases (grants P60-AR48098, R01-
AR055287, R01-AR054155, and R21-AR059412) and the Falk
Medical Research Trust. The Osteoarthritis Initiative (OAI) is
a public–private partnership comprised of 5 contracts (N01-
AR-2-2258, N01-AR-2-2259, N01-AR-2-2260, N01-AR-2-2261,
and N01-AR-2-2262) funded by the NIH and conducted by the
OAI Study Investigators. Private funding partners include
Merck Research Laboratories, Novartis Pharmaceuticals,
GlaxoSmithKline, and Pfizer. Private sector funding for the
OAI is managed by the Foundation for the NIH.
1Jungwha Lee, PhD, MPH, Jing Song, MS, Pamela A.
Semanik, PhD, Rowland W. Chang, MD, MPH, Leena
Sharma, MD, Linda van Horn, PhD, RD, Dorothy D. Dunlop,
PhD: Northwestern University, Chicago, Illinois;2Jennifer M.
Hootman, PhD: Centers for Disease Control and Prevention,
Atlanta, Georgia;3Joan M. Bathon, MD: Columbia University,
New York, New York;4Charles B. Eaton, MD: Brown Univer-
sity, Pawtucket, Rhode Island;5Marc C. Hochberg, MD: Uni-
versity of Maryland, Baltimore;
W. Jerry Mysiw, MD: Ohio State University, Columbus;
7C. Kent Kwoh, MD: University of Pittsburgh, Pittsburgh, Penn-
sylvania;8Michael Nevitt, PhD: University of California at San
Dr. Eaton has received honoraria (less than $10,000) from
the Medical University of South Carolina.
Address correspondence to Jungwha Lee, PhD, MPH, As-
sistant Professor, Northwestern University, 680 North Lake
Shore Drive, Suite 1400, Chicago, IL 60611. E-mail: jung
Submitted for publication February 1, 2012; accepted in
revised form May 25, 2012.
6Rebecca Jackson, MD,
Arthritis Care & Research
Vol. 65, No. 1, January 2013, pp 53–61
© 2013, American College of Rheumatology
SPECIAL THEME ARTICLE: OBESITY AND THE RHEUMATIC DISEASES
activity programs can reduce pain, improve physical per-
formance, reduce depressive symptoms, and prevent or
delay disability in knee OA (5–8). In addition to disease-
specific benefits, randomized clinical trials show that
physical activity can improve muscle strength and in-
crease aerobic capacity, flexibility, and strength (9–11).
Recent federal guidelines now include people with arthri-
tis in the physical activity recommendations to promote
these benefits (12).
Despite important health benefits from being physically
active, persons with arthritis are particularly inactive and
are at risk for poor health outcomes (13,14). In a national
US survey, 44% of persons with arthritis were classified as
inactive (i.e., reporting no sustained 10-minute periods of
moderate-to-vigorous [MV] physical activity in a week)
compared to 36% of adults without arthritis (15). Physical
inactivity may account for an estimated 21% of activity
limitations attributed to arthritis (16). Furthermore, phys-
ical inactivity threatens full participation in both employ-
ment opportunities and independent community living, as
well as leads to increased health care costs (17).
Identifying predictors of inactivity is important to de-
velop public health interventions aimed at reducing inac-
tivity. Therefore, this study simultaneously investigated
risk factors that are modifiable and related to inactivity to
identify strategic targets for public health intervention. For
example, knee pain and function are commonly viewed as
barriers to being physically active for adults with knee OA,
and significant associations between inactivity were re-
ported in the National Health Interview Survey (NHIS)
(15) and the Canadian National Population Health Survey
(NPHS) (18) studies. Similarly, being overweight, depres-
sive symptoms, and smoking are associated with inactivity
(7,19,20), and knee confidence is implicated by its strong
association with physical function (21). Literature from
the general population shows that low fiber intake has
been associated with low physical activity levels (22).
There are limited reports on objectively measured phys-
ical activity that examine characteristics of inactivity. The
purpose of this study was to identify modifiable risk fac-
tors that may increase the frequency of physical inactivity
among adults with knee OA and to calculate the attribut-
able fraction (AF) of modifiable risk factors that account
for the excess physical inactivity in this sample.
PATIENTS AND METHODS
Study population. This physical activity study evalu-
ated a subcohort from the Osteoarthritis Initiative (OAI),
a prospective study investigating risk factors and bio-
markers associated with the development and progression
of knee OA in adults ages 45–79 years at enrollment, with
or at high risk to develop knee OA. Annual OAI evalua-
tions began in 2004 at 4 clinical sites (Baltimore, Mary-
land; Columbus, Ohio; Pittsburgh, Pennsylvania; and Paw-
tucket, Rhode Island) and are currently ongoing (http://
www.oai.ucsf.edu/datarelease/About.asp). Internal review
board approval was obtained at each of the participating
sites. Each participant provided written informed consent.
The OAI excluded individuals with rheumatoid or inflam-
matory arthritis; severe joint space narrowing in both
knees on the baseline knee radiograph or unilateral total
knee replacement and severe joint space narrowing in the
other knee; bilateral total knee replacement or plans to
have bilateral knee replacement in the next 3 years; inabil-
ity to undergo a 3.0T magnetic resonance imaging exami-
nation of the knee because of contraindications; positive
pregnancy test; inability to provide a blood sample; use of
ambulatory aides other than a single straight cane for more
than 50% of the time during ambulation; comorbid con-
ditions that might interfere with the ability to participate
in a 4-year study; or current participation in a double-
blind randomized trial. All OAI participants underwent
knee radiography at baseline. The radiographic acquisition
protocol may be found at http://www.oai.ucsf.edu/data
release/OperationsManuals.asp. Baseline films were as-
sessed by clinic readers for the Osteoarthritis Research
Society International (OARSI) atlas grades of tibiofemoral
osteophytes and joint space narrowing (23). The baseline
visit identified 2,679 participants with radiographic evi-
dence of knee OA (i.e., definite or osteophyte grade of ?1,
according to the OARSI atlas) in one or both knees from
the total OAI enrollment of 4,796 persons in 2004–2005.
Accelerometer monitoring of physical activity study was
offered to a subcohort of OAI participants at the 48-month
followup visit between 2008 and 2010. A total of 2,127
persons consented to participate in accelerometer moni-
toring, representing 78.4% of the 2,712 eligible subcohort
participants (another 1,543 OAI participants had visits
that preceded the accelerometer study start date and 541
were deceased or withdrew from the OAI study or did not
return for the 48-month visit). This report included only
the 1,223 participants with baseline radiographic knee OA
as shown in Figure 1. Accelerometer and 48-month visit
data were merged with OAI public data (24) containing
information on participant characteristics.
Outcome: physical activity measures. Physical activity
was measured using a GT1M Actigraph accelerometer, a
small uniaxial accelerometer that measures vertical accel-
erations (25). Uniaxial accelerometer validation studies
Significance & Innovations
● We investigated the potential public health impact
of modifiable factors including being obese/
overweight, low-fiber diet, inadequate dietary fat,
smoking status, depressive symptoms, severe knee
pain, and severe knee dysfunction related to phys-
ical inactivity in adults with knee osteoarthritis.
● The public health importance of each modifiable
risk factor on physical inactivity was estimated
using the attributable fraction.
● Modifiable factors significantly associated with
physical inactivity (being obese/overweight, the
quality of the diet, severe knee pain, and severe
knee dysfunction) should be considered in devel-
oping physical activity interventions that target
arthritis populations with low activity levels.
54Lee et al
against whole-body indirect calorimetry showed high cor-
relation with metabolic equivalent (r ? 0.93) and total
energy expenditure (r ? 0.93) (26). The accuracy (27) and
test–retest reliability (28) of Actigraph accelerometers un-
der field conditions are established in many populations,
including persons with OA (29). Accelerometers output
an activity count, which is the weighted sum of the num-
ber of accelerations measured over 1-minute periods,
where the weights are proportional to the magnitude of
Trained research personnel gave participants uniform
scripted instructions to wear the unit on a belt at the
natural waistline on the right hip in line with the right
axilla upon arising in the morning and continuously until
retiring at night, except during water activities, for 7 con-
secutive days. Participants maintained a daily log to re-
cord time spent in water and cycling activities, which may
not be fully captured by accelerometers. Accelerometer
data were analytically filtered using methodology vali-
dated in adults with rheumatic disease (30–32). Periods of
non-wear were defined as ?90 minutes with zero activity
counts (allowing for 2 interrupted minutes with counts
Accelerometer data included ?4 valid days for each
participant. A valid day of monitoring was defined as ?10
wear hours in a 24-hour period (30). Total daily minutes of
MV physical activity were calculated using methodology
from the National Institutes of Health (counts ?2,020/
minute occurring in bouts lasting ?10 minutes, with al-
lowance for interruptions of 1 or 2 minutes below the MV
threshold) (30). Each person was classified according to
the US Department of Health and Human Services (DHHS)
physical activity guidelines as meets recommendations
(?150 minutes of MV activity bouts/week in bouts lasting
at least 10 minutes), low active (10–149 minutes of MV
bouts/week), or inactive (zero minutes of MV activity
Modifiable risk factors. Modifiable general health fac-
tors included body mass index (BMI) and depressive
symptoms. BMI was calculated from measured height (m2)
and weight (kg). Persons were classified as normal weight
(BMI 18.5–24.9 kg/m2), overweight (BMI 25.0–29.9 kg/m2),
or obese (BMI ?30 kg/m2). High depressive symptoms
were assessed by a score ?16 from the Center for Epide-
miologic Studies Depression Scale (CES-D) (33). Knee-
specific factors included knee function, knee pain, and
knee confidence. Self-reported current knee function and
pain in the past 7 days were obtained from the Western
Ontario and McMaster Universities Osteoarthritis Index
(WOMAC), Likert version 3.1, modified in the OAI (34).
Person-level scores used the maximum WOMAC value of
the 2 knees. For the purpose of analysis, we classified
WOMAC knee function symptoms based on tertiles as
follows: no dysfunction (bottom tertile) ? 0; moderate
dysfunction ? 0.1–11.2; and severe dysfunction (top ter-
tile) ? 11.3–68. Similarly, WOMAC knee pain was classi-
fied using tertiles as no pain (bottom tertile) ? 0; moderate
pain ? 0.1–3.9; and severe pain (top tertile) ? 4–20. Knee
confidence was assessed using the Knee Injury and Osteo-
arthritis Outcome Score (KOOS) (35) question, “How much
are you troubled with lack of confidence in your knees?”
with possible responses of not at all, mildly, moderately,
severely, and extremely.
Modifiable behavior factors included smoking status
and dietary intake. Smoking status was dichotomized as
current smoking versus not current smoking or missing.
Dietary intake was assessed at baseline using the Block
2000 systematic nutrition assessment (36,37). For dietary
intake variables, we focused on fat intake and dietary fiber
intake as they represent the 2 extremes of low- and high-
calorie density of food. Dietary factors were dichotomized
as high fat (?35% of daily calories) versus adequate/low
fat intake, and as inadequate (?20 gm/day) versus ade-
quate fiber intake (38,39). Additional nutritional factors
measured by the Block 2000 were only weakly associated
with inactivity status and therefore were not considered
relevant as modifiable factors. The questionnaires can be
obtained from the publically available OAI web site
Descriptive factors. Covariates were measured at the
OAI 48-month visit except where noted. Descriptive cova-
riates included sociodemographics and prior health fac-
tors. Sociodemographic baseline factors included race,
age, sex, living status, education, and employment (at 24-
month visit). Descriptive prior health factors included fre-
quent knee symptoms, comorbidity, and the report of a
total knee replacement prior to enrollment. Frequent knee
symptoms were ascertained from a positive response to the
question, “During the past 12 months, have you had pain,
aching, or stiffness in or around your right/left knee on
most days for at least one month?” Comorbidities were
Initiative (OAI) cohort
n = 4,796
Eligible subcohort for
accelerometer study at
OAI 48-month visit
n = 2,712
Not eligible (OAI 48-month
visit prior to start date:
1,543; did not return/
n = 2,084
n = 2,127
Not participating in
n = 585
participants with baseline
radiographic knee OA
n = 1,223
baseline radiographic knee
n = 904
Figure 1. Flow chart of analytical sample OAI participants in
accelerometer study. OA ? osteoarthritis.
Modifiable Risk Factors for Physical Inactivity in Knee OA55
Table 1. Physical activity distribution of adults with radiographic knee OA (n ? 1,089)*
No. Inactive, %Low active, %Meets reccomendations, %
P for trend†
High-school or less
Chronic knee pain
Total knee replacement prior to
General health factors
BMI normal (18.5–24.9 kg/m2)‡
BMI overweight (25.0–29.9 kg/m2)
BMI obese (?30 kg/m2)
High (CES-D ?16)
Dietary fiber intake
Knee-specific health factors
Moderate (middle tertile)
Severe (top tertile)
WOMAC knee pain¶
Moderate (middle tertile)
Severe (top tertile)
* OA ? osteoarthritis; BMI ? body mass index; CES-D ? Center for Epidemiologic Studies Depression Scale; WOMAC ? Western Ontario and
McMaster Universities Osteoarthritis Index.
† Mantel-Haenszel chi-square test for trend (1df) except for race, sex, and living arrangement comparisons, which used chi-square test for overall
‡ Includes 1 person with BMI below the normal range.
§ WOMAC function tertiles based on maximum of left and right knees: no dysfunction (bottom tertile) ? 0; moderate dysfunction ? 0.1–11.2; and
severe dysfunction (top tertile) ? 11.3–68.
¶ WOMAC pain tertiles based on maximum of left and right knees: no pain (bottom tertile) ? 0; moderate pain ? 0.1–3.9; and severe pain (top tertile) ?
56Lee et al
assessed by the modified Charlson comorbidity index (40)
at the 24-month visit.
Statistical analysis. Univariate analyses of baseline trend
effects across physical activity levels were evaluated by a
Mantel-Haenszel test for ordinal categories. A chi-square
test for overall differences was applied to nominal vari-
ables. Modifiable factors associated with physical inactiv-
ity were evaluated by logistic regression, controlling for
all descriptive and modifiable factors; an associated 95%
confidence interval (95% CI) that falls above 1 indicates a
significant association. Further analyses added interaction
terms between sex and each modifiable factor to logistic
The AF related to inactivity was estimated for each
modifiable risk factor (41). The sample AF estimates “ex-
cess” inactivity based on the risk factor frequency and its
association with inactivity. The term “excess” conceptu-
ally refers to the reduction in the outcome that would
occur if the risk factor were removed from the population.
In a cross-sectional sample, the AF is the potential reduc-
tion in the outcome (e.g., inactivity) if the risk factor was
totally absent (e.g., no obesity) (42,43), but it does not
imply cause and effect. An average AF (AAF) accounts for
individuals with multiple factors, estimating the excess
proportion of the outcome that can be attributed to any of
the designated risk factors (44). The AAF is usually lower
than the sum of individual AFs because an outcome is
typically attributable to more than one risk factor. The AF
estimates were assessed using Poisson regression with ro-
bust error variance and the sample prevalence of the mod-
ifiable risk factors employing SAS, version 9.2, with SAS
macro in 2010–2011 (45,46).
Table 2. Modifiable health factor ORs for physical inactivity in adults with radiographic knee OA (n ? 1,089)*
Modifiable health factors No. Inactive, %
OR (95% CI)
OR (95% CI)†
OR (95% CI)‡
Dietary fiber intake
General health factors
Moderate (middle tertile)
Severe (top tertile)
WOMAC knee pain#
Moderate (middle tertile)
Severe (top tertile)
* ORs ? odds ratios; OA ? osteoarthritis; 95% CI ? 95% confidence interval; BMI ? body mass index; WOMAC ? Western Ontario and McMaster
Universities Osteoarthritis Index.
† OR adjusted for descriptive factors (age, sex, race, living status, education, employment, chronic knee pain, total knee replacement, and comorbidi-
ties) with associated 95% CI.
‡ OR adjusted for all descriptive and other modifiable factors with associated 95% CI.
§ Includes 1 person with BMI below the normal range. BMI ranges are 18.5–24.9 kg/m2for normal, 25.0–29.9 kg/m2for overweight, and ?30 kg/m2
¶ WOMAC function tertiles based on maximum of left and right knees: no dysfunction (bottom tertile) ? 0; moderate dysfunction ? 0.1–11.2; and
severe dysfunction (top tertile) ? 11.3–68.
# WOMAC pain tertiles based on maximum of left and right knees: no pain (bottom tertile) ? 0; moderate pain ? 0.1–3.9; and severe pain (top tertile) ? 4–20.
Modifiable Risk Factors for Physical Inactivity in Knee OA 57
A total of 1,223 persons, ages 49–84 years, with radio-
graphic knee OA participated in physical activity mea-
surement using accelerometers at the 48-month OAI visit.
At the baseline OAI visit, 67.1% of these had definite joint
space narrowing equivalent of a Kellgren/Lawrence grade
of 3 or 4 (47). Participants in this study had similar base-
line age (62.0 years versus 62.8 years) and BMI (29.2 kg/m2
versus 29.8 kg/m2) compared to the nonparticipating OAI
radiographic knee OA cohort, but were more frequently
male (44.9% versus 38.9%) and white (81.0% versus 75.3%),
with slightly less pain (mean WOMAC pain score 3.5
versus 4.6). Of the participants, 1,111 (90.8%) had at least
4 valid days of accelerometer monitoring data; a total of
1,089 participants (all but 2%) had complete covariate
These 1,089 adults with knee OA had a mean age of
66.1 years, were primarily female (54.8%), white (83.7%),
working (53.9%), did not live alone (78.3%), and had post
high-school education (84.9%). Few smoked (5.7%) or
had high depressive symptoms (11.5%), but substantial
portions reported high fat intake (46.3%), inadequate di-
etary fiber (79.2%), some limitations in knee function
(69.1%), knee pain (68.6%), some trouble with lack of
knee confidence (53.2%), and were overweight/obese
(79.0%). In regard to physical activity, 48.9% (39.8% of
men and 56.3% of women) were inactive, demonstrating
no 10-minute bouts of MV activity during the week of
monitoring. Only 10.2% attained recommended activity
Descriptive and modifiable characteristics of this sample
by physical activity characteristics are shown in Table 1.
All descriptive factors were significantly associated with
physical activity levels. Among modifiable factors, greater
levels of physical inactivity were significantly associated
with greater prevalence of overweight/obesity, inadequate
dietary fiber intake, greater WOMAC knee dysfunction, and
greater WOMAC knee pain.
The association of modifiable factors with inactivity
quantified as odds ratios (ORs) is summarized in Table 2.
Modifiable factors significantly related to inactivity were
being overweight (OR 1.8, 95% CI 1.2–2.5) or obese (OR
3.9, 95% CI 2.6–5.7), inadequate dietary fiber intake (OR
1.6, 95% CI 1.2–2.2), severe dysfunction (OR 1.9, 95% CI
1.3–2.8), and severe pain (OR 1.7, 95% CI 1.1–2.5) after
controlling for descriptive factors. Being overweight (OR
1.8, 95% CI 1.2–2.5) or obese (OR 3.7, 95% CI 2.5–5.4) and
inadequate dietary fiber intake (OR 1.5, 95% CI 1.1–2.1)
were significantly related to inactivity when controlling
for all descriptive and modifiable factors.
The influence of modifiable factors on inactivity is fur-
ther examined from a public health perspective. Figure 2
presents the modifiable factor AAFs adjusted for descrip-
tive and other modifiable factors. The AAF accounts for
the fact that some individuals have multiple modifiable
risk factors. For example, among overweight/obese adults
in this study, 80.2% had low-fiber diets and 70.8% had
knee pain. The AAF model includes the significant mod-
ifiable factors based on the bivariate relationship between
inactivity using logistic regression, i.e., dietary fiber in-
take, weight status, knee function, and knee pain. Being
overweight/obese had a statistically significant 23.8%
(95% CI 10.5–38.6%) relationship to excess inactivity.
Inadequate dietary fiber was significantly related to an-
other 12.1% (95% CI 0.1–24.5%). Together, being obese/
overweight and having a low-fiber diet accounted for
35.9% of excess inactivity in these adults with knee OA.
Remaining modifiable factors were related to ?7% excess
inactivity and were not significant.
This study contributes to public health efforts to improve
health outcomes of persons with arthritis by examining
the association of potentially modifiable risk factors asso-
ciated with physical inactivity in a large cohort of adults
with radiographic evidence of knee OA. An important
strength of this study is the objective assessment of phys-
ical inactivity. Only 1 of 10 adults with knee OA had
recommended physical activity levels. Notably, more than
one-third of men and half the women were completely
inactive, doing no sustained MV activity that lasted ?10
minutes in a week. Modifiable factors significantly associ-
ated with inactivity were being overweight/obese and con-
suming a diet with inadequate fiber, the report of severe
knee limitations, and severe knee pain. More than 23.8%
of excess inactivity was related to being overweight/obese,
Figure 2. Adjusted average attributable fraction of inactivity for
modifiable factors; adjusted for descriptive factors (age, sex, race,
living status, education, employment, chronic knee pain, total
knee replacement, comorbidities) and all modifiable factors
(obese/overweight, inadequate dietary fiber, knee dysfunction,
knee pain, high-fat diet, smoking, high depressive symptoms, or
being troubled by knee confidence). ?? ? aggregated contribution
of remaining modifiable risk factors (high-fat diet, smoking, high
depressive symptoms, or being troubled by knee confidence).
58Lee et al
and another 12.1% was related to inadequate dietary fiber
consumption after accounting for other descriptive and
A low level of physical activity among adults with ar-
thritis is a recognized public health concern. However,
assessing the magnitude of the problem has been a chal-
lenge due to differing methods for assessing physical ac-
tivity. Earlier studies that relied on self-reported physical
activity levels estimate that 23.8–57.8% of adults with
arthritis in the US are inactive (13,15,16,18,19,48). Imbed-
ded into these estimates are differences related to the
self-report of physical activity and how inactivity was
defined. Inactivity was defined by no reported leisure time
activity (19,48), ?10 minutes/week MV leisure activities
(13), ?3 sessions/month lasting ?15 minutes of activities
associated with moderate intensity energy expenditure
(16,18), and no reported activities lasting ?10 minutes
(15). In this study, the definition of physical inactivity is
anchored on the federal DHHS definition and is assessed
by objective accelerometer monitoring.
Modifiable factors were evaluated from 2 perspectives.
The first perspective identifies factors associated with
physical inactivity at the level of the individual. Modifi-
able factors significantly associated with inactivity based
on adjusted ORs were obesity, knee pain, knee dysfunc-
tion, and dietary fiber intake. Earlier studies on adults
reporting an arthritis diagnosis (18,19) found a significant
relationship between inactivity and being overweight, but
a 2002 NHIS study did not find a significant association
(15). These reports evaluate broader arthritis populations
than the current study and rely on self-report to assess
inactivity and weight status. Self-report is related to un-
derreporting of weight (49) and overreporting of physical
activity amount/intensity (50). It is not known how report-
ing accuracy may influence the apparent association be-
tween inactivity and weight status, but measurement is-
sues may contribute to the mixed findings across earlier
Significant associations between inactivity and self-
reported knee dysfunction (OR 1.9) and pain (OR 1.7)
were in line with findings from the NHIS (15) and Cana-
dian NPHS (18) studies. We also found a significant asso-
ciation between inactivity and inadequate fiber intake
(OR 1.6). In the general adult population, low fiber intake
has been associated with low physical activity levels
(22,51). This association may partially reflect the associa-
tion of a low-fiber diet, representing low intake of fruits,
vegetables, whole grains, and high consumption of refined
carbohydrates/sugars related to snacking and a sedentary
lifestyle (52,53). Thus, a low-fiber diet may be a marker for
unhealthy behaviors that include inactivity and high fat/
high sugar snacking.
A second public health perspective examined the influ-
ence of each modifiable factor on inactivity by estimating
the AAF for the sample. The sample AAF has public
health relevance because the metric incorporates popula-
tion criteria related to the risk factor prevalence plus its
association with the outcome. Recognizing that many in-
dividuals had multiple modifiable risk factors (e.g., 80.2%
of overweight/obese adults had low-fiber diets and 70.8%
reported pain), we examined the simultaneous effect of all
modifiable risk factors on inactivity. While pain (AAF
6.2%) and dysfunction (AAF 2.6%) are associated with
lower levels of inactivity, being obese/overweight (AAF
23.8%) and inadequate dietary fiber (AAF 12.1%) explain
a significant and larger proportion of inactivity. The re-
sults reinforce the contribution of excess weight and poor
Pain and poor function are commonly viewed as bar-
riers to being physically active for adults with knee OA
(15,18,54). These findings indicate that being overweight/
obese and an unhealthy diet are also important to con-
sider. There is evidence that higher BMI is related to
greater knee pain and poor function in adults with knee
OA (55–58). In turn, high levels of pain are associated with
binge eating (59). If obesity due to poor dietary patterns
contributes to knee pain and resulting poor function
through mechanical stress due to excess weight on the
joint, then obesity supported by poor dietary choices may
contribute to the relationship between knee pain and in-
activity. However, randomized controlled trials show that
exercise is safe and effective for overweight/obese adults
with OA (58). Taken together, these results support incor-
porating weight loss and diet modification into interven-
tions designed to promote health benefits from physical
There are limitations to acknowledge in the present
study. Accelerometers do not provide information on the
type of the physical activity (e.g., household, transporta-
tion), information which may be helpful in targeting inter-
ventions. The accelerometer used in this study cannot
capture water activities and may underestimate upper
body movement or activities such as cycling. Diary infor-
mation showed a median of 0 minutes/day spent in water
and cycling activities, so the potential underestimate of
inactivity is negligible. Radiographic data on knee joint
damage and dietary information were only available from
baseline, 4 years prior to the current study. Because joint
damage does not improve over time and people with sub-
sequent knee replacements were excluded, radiographic
verification remains valid. Dietary fiber intake tends to
remain stable or decrease slightly over a 3- to 4-year pe-
riod, as demonstrated by control groups from large nutri-
tional trials (60,61); a potential underestimate of inade-
quate fiber intake would likely understate the strength of
its relationship with inactivity found in our analyses. Self-
reported data are commonly subject to social desirability
bias and recall bias. However, to minimize biases, we have
used validated questionnaires such as Block 2000 for di-
etary variables, the CES-D for depressive symptoms, the
WOMAC for knee function and pain, and the KOOS for
knee confidence. Our results will be strengthened if par-
ticipants overreported dietary fiber intake as we would
have underestimated the effects of dietary fiber. Finally,
causality cannot be inferred from these cross-sectional
data. These limitations must be balanced against the sub-
stantial strengths of this study, which include the large
sample size, clinical measures of height and weight as
opposed to self-reports, radiographic verification of knee
OA, and the age and sex diversity of this OA cohort. An
important strength of this study is that the federal DHHS
Modifiable Risk Factors for Physical Inactivity in Knee OA59
definition of inactivity (12) based on objective accelerom-
eter monitoring was used.
Despite substantial health benefits related to physical
activity, adults with radiographic knee OA were particu-
larly inactive. A substantial 48.9% of adults with knee
OA were classified as inactive, demonstrating no 10-
minute episodes of MV activity in a week. There is a
critical need to intensify public health efforts to reduce
physical inactivity among adults with knee OA. Being
obese/overweight, the quality of the diet, severe pain,
severe dysfunction, and levels of physical activity are in-
terrelated in adults with knee OA. One cannot hope to
improve physical activity patterns in adults with knee
OA without consideration for weight management, diet,
and OA pain and disability, as all may affect successful
achievement of activity goals. All components should be
considered in developing physical activity interventions
that target arthritis populations with low activity levels.
All authors were involved in drafting the article or revising it
critically for important intellectual content, and all authors ap-
proved the final version to be submitted for publication. Dr. Lee
had full access to all of the data in the study and takes responsi-
bility for the integrity of the data and the accuracy of the data
Study conception and design. Lee, Song, Chang, van Horn,
Bathon, Jackson, Mysiw, Nevitt, Dunlop.
Acquisition of data. Lee, Song, Bathon, Eaton, Hochberg, Jackson,
Kwoh, Mysiw, Nevitt.
Analysis and interpretation of data. Lee, Song, Hootman,
Semanik, Chang, Sharma, van Horn, Bathon, Eaton, Jackson,
Kwoh, Nevitt, Dunlop.
ROLE OF THE STUDY SPONSOR
Merck Research Laboratories, Novartis Pharmaceuticals, Glaxo-
SmithKline, and Pfizer had no role in the study design or in the
collection, analysis, or interpretation of the data, the writing of the
manuscript, or the decision to submit the manuscript for publi-
cation. Publication of this article was not contingent upon ap-
proval by Merck Research Laboratories, Novartis Pharmaceuti-
cals, GlaxoSmithKline, and Pfizer.
1. Centers for Disease Control and Prevention. Prevalence of
doctor-diagnosed arthritis and arthritis-attributable activity
limitation: United States, 2007-2009. MMWR Morb Mortal
Wkly Rep 2010;59:1261–5.
2. Lawrence RC, Felson DT, Helmick CG, Arnold LM, Choi H,
Deyo RA, et al. Estimates of the prevalence of arthritis and
other rheumatic conditions in the United States. Part II. Ar-
thritis Rheum 2008;58:26–35.
3. Yelin E. Cost of musculoskeletal diseases: impact of work
disability and functional decline. J Rheumatol Suppl 2003;68:
4. Felson DT, Lawrence RC, Dieppe PA, Hirsch R, Helmick CG,
Jordan JM, et al. Osteoarthritis: new insights. Part 1: the dis-
ease and its risk factors. Ann Intern Med 2000;133:635–46.
5. Deyle GD, Henderson NE, Matekel RL, Ryder MG, Garber MB,
Allison SC. Effectiveness of manual physical therapy and
exercise in osteoarthritis of the knee: a randomized controlled
trial. Ann Intern Med 2000;132:173–81.
6. Ettinger WH Jr, Burns R, Messier SP, Applegate W, Rejeski WJ,
Morgan T, et al. A randomized trial comparing aerobic exer-
cise and resistance exercise with a health education program
in older adults with knee osteoarthritis. The Fitness Arthritis
and Seniors Trial (FAST). JAMA 1997;277:25–31.
7. Penninx BW, Rejeski WJ, Pandya J, Miller ME, Di Bari M,
Applegate WB, et al. Exercise and depressive symptoms: a
comparison of aerobic and resistance exercise effects on emo-
tional and physical function in older persons with high and
low depressive symptomatology. J Gerontol B Psychol Sci Soc
8. Penninx BW, Messier SP, Rejeski WJ, Williamson JD, DiBari
M, Cavazzini C, et al. Physical exercise and the prevention of
disability in activities of daily living in older persons with
osteoarthritis. Arch Intern Med 2001;161:2309–16.
9. Ades PA, Ballor DL, Ashikaga T, Utton JL, Nair KS. Weight
training improves walking endurance in healthy elderly per-
sons. Ann Intern Med 1996;124:568–72.
10. Cress ME, Buchner DM, Questad KA, Esselman PC, deLateur
BJ, Schwartz RS. Exercise: effects on physical functional per-
formance in independent older adults. J Gerontol A Biol Sci
Med Sci 1999;54:242–8.
11. Fiatarone MA, Marks EC, Ryan ND, Meredith CN, Lipsitz LA,
Evans WJ. High-intensity strength training in nonagenarians:
effects on skeletal muscle. JAMA 1990;263:3029–34.
12. US Department of Health and Human Services. 2008 physical
activity guidelines for Americans. 2008. URL: http://www.
13. Fontaine KR, Heo M, Bathon J. Are US adults with arthritis
meeting public health recommendations for physical activity?
Arthritis Rheum 2004;50:624–8.
14. Keysor JJ. Does late-life physical activity or exercise prevent
or minimize disablement? A critical review of the scientific
evidence. Am J Prev Med 2003;25:129–36.
15. Shih M, Hootman JM, Kruger J, Helmick CG. Physical activity
in men and women with arthritis National Health Interview
Survey, 2002. Am J Prev Med 2006;30:385–93.
16. Badley EM, Ansari H. Arthritis and arthritis-attributable ac-
tivity limitations in the United States and Canada: a cross-
border comparison. Arthritis Care Res (Hoboken) 2010;62:
17. Wang G, Helmick CG, Macera C, Zhang P, Pratt M. Inactivity-
associated medical costs among US adults with arthritis. Ar-
thritis Rheum 2001;45:439–45.
18. Kaplan MS, Huguet N, Newsom JT, McFarland BH. Charac-
teristics of physically inactive older adults with arthritis:
results of a population-based study. Prev Med 2003;37:61–7.
19. Mehrotra C, Chudy N, Thomas V. Obesity and physical inac-
tivity among Wisconsin adults with arthritis. WMJ 2003;102:
20. Strine TW, Balluz LS, Ford ES. The associations between
smoking, physical inactivity, obesity, and asthma severity in
the general US population. J Asthma 2007;44:651–8.
21. Colbert CJ, Song J, Dunlop D, Chmiel JS, Hayes KW, Cahue S,
et al. Knee confidence as it relates to physical function out-
come in persons with or at higher risk for knee osteoarthritis
in the Osteoarthritis Initiative. Arthritis Rheum 2012;64:
22. Gillman MW, Pinto BM, Tennstedt S, Glanz K, Marcus B,
Friedman RH. Relationships of physical activity with dietary
behaviors among adults. Prev Med 2001;32:295–301.
23. Altman RD, Hochberg M, Murphy WA Jr, Wolfe F, Lequesne
M. Atlas of individual radiographic features in osteoarthritis.
Osteoarthritis Cartilage 1995;Suppl A:3–70.
24. The Osteoarthritis Initiative. Osteoarthritis Initiative (OAI)
database. In: Center OIC, ed. Baseline 0.2.2, 12 month visit
1.2.1, 24 month visit 3.2.1 and 36 month visit 5.2.1 ed. San
Francisco, CA; 2010. URL: http://www.oai.ucsf.edu.
25. Matthews CE, Ainsworth BE, Thompson RW, Bassett DR Jr.
Sources of variance in daily physical activity levels as mea-
sured by an accelerometer. Med Sci Sports Exerc 2002;34:
26. Kumahara H, Schutz Y, Ayabe M, Yoshioka M, Yoshitake Y,
Shindo M, et al. The use of uniaxial accelerometry for the
assessment of physical-activity–related energy expenditure: a
validation study against whole-body indirect calorimetry. Brit
J Nutr 2004;91:235–43.
60Lee et al
27. Brage S, Wedderkopp N, Franks PW, Andersen LB, Froberg K. Download full-text
Reexamination of validity and reliability of the CSA monitor
in walking and running. Med Sci Sports Exerc 2003;35:
28. Welk GJ, Schaben JA, Morrow JR Jr. Reliability of accelerom-
etry-based activity monitors: a generalizability study. Med Sci
Sports Exerc 2004;36:1637–45.
29. Farr JN, Going SB, Lohman TG, Rankin L, Kasle S, Cornett M,
et al. Physical activity levels in patients with early knee
osteoarthritis measured by accelerometry. Arthritis Rheum
30. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T,
McDowell M. Physical activity in the United States measured
by accelerometer. Med Sci Sports Exerc 2008;40:181–8.
31. Song J, Semanik P, Sharma L, Chang RW, Hochberg MC,
Mysiw WJ, et al. Assessing physical activity in persons with
knee osteoarthritis using accelerometers: data in the osteo-
arthritis initiative. Arthritis Care Res (Hoboken) 2010;62:
32. Semanik P, Song J, Chang RW, Manheim L, Ainsworth B,
Dunlop D. Assessing physical activity in persons with rheu-
matoid arthritis using accelerometry. Med Sci Sports Exerc
33. Pandya R, Metz L, Patten SB. Predictive value of the CES-D in
detecting depression among candidates for disease-modifying
multiple sclerosis treatment. Psychosomatics 2005;46:131–4.
34. McGrory BJ, Harris WH. Can the Western Ontario and
McMaster Universities (WOMAC) Osteoarthritis Index be
used to evaluate different hip joints in the same patient?
J Arthroplasty 1996;11:841–4.
35. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD.
Knee Injury and Osteoarthritis Outcome Score (KOOS): de-
velopment of a self-administered outcome measure. J Orthop
Sports Phys Ther 1998;28:88–96.
36. Block G, Woods M, Potosky A, Clifford C. Validation of a
self-administered diet history questionnaire using multiple
diet records. J Clin Epidemiol 1990;43:1327–35.
37. Subar AF, Thompson FE, Kipnis V, Midthune D, Hurwitz P,
McNutt S, et al. Comparative validation of the Block, Willett,
and National Cancer Institute food frequency questionnaires:
the Eating at America’s Table Study. Am J Epidemiol 2001;
38. Dietary Guidelines Advisory Committee on the Dietary
Guidelines for Americans. Report on the dietary guidelines
for Americans. Washington (DC): US Department of Health
and Human Services; 2010.
39. Institute of Medicine Panel on Macronutrients, Subcommit-
tees on Upper Reference Levels of Nutrients and Interpreta-
tion and Uses of Dietary Reference Intakes, Standing Commit-
tee on the Scientific Evaluation of Dietary Reference Intakes.
Dietary reference intakes for energy, carbohydrates, fiber, fat,
fatty acids, cholesterol, protein, and amino acids (macronu-
trients). Washington (DC): National Academies Press; 2005.
40. Katz JN, Chang LC, Sangha O, Fossel AH, Bates DW. Can
comorbidity be measured by questionnaire rather than med-
ical record review? Med Care 1996;34:73–84.
41. Levin ML. The occurrence of lung cancer in man. Acta Unio
Internationalis Contra Cancrum 1953;19:531–41.
42. Fleiss JL. Attributable risk. In: Statistical methods for rates
and proportions. 2nd ed. New York: John Wiley & Sons; 1981.
43. Benichou J. A review of adjusted estimators of attributable
risk. Stat Methods Med Res 2001;10:195–216.
44. Eide GE, Gefeller O. Sequential and average attributable frac-
tions as aids in the selection of preventive strategies. J Clin
45. Spiegelman D, Hertzmark E, Wand HC. Point and interval
estimates of partial population attributable risks in cohort
studies: examples and software. Cancer Causes Control 2007;
46. Hertzmark E. Software for computing full and partial popu-
lation attributable risks and their confidence intervals. URL:
47. Kellgren JH, Lawrence JS. Radiological assessment of osteo-
arthrosis. Ann Rheum Dis 1957;16:494–502.
48. Hootman JM, Macera CA, Ham SA, Helmick CG, Sniezek JE.
Physical activity levels among the general US adult popula-
tion and in adults with and without arthritis. Arthritis Rheum
49. Villanueva EV. The validity of self-reported weight in US
adults: a population based cross-sectional study. BMC Public
50. Irwin ML, Ainsworth BE, Conway JM. Estimation of energy
expenditure from physical activity measures: determinants of
accuracy. Obes Res 2001;9:517–25.
51. Eaton CB, McPhillips JB, Gans KM, Garber CE, Assaf AR,
Lasater TM, et al. Cross-sectional relationship between diet
and physical activity in two southeastern New England com-
munities. Am J Prev Med 1995;11:238–44.
52. Spruijt-Metz D, Belcher B, Anderson D, Lane CJ, Chou CP,
Salter-Venzon D, et al. A high-sugar/low-fiber meal compared
with a low-sugar/high-fiber meal leads to higher leptin and
physical activity levels in overweight Latina females. J Am
Diet Assoc 2009;109:1058–63.
53. Thayer R. Calm energy: how people regulate mood with food
and exercise. New York: Oxford University Press; 2001.
54. McBeth J, Nicholl BI, Cordingley L, Davies KA, Macfarlane
GJ. Chronic widespread pain predicts physical inactivity: re-
sults from the prospective EPIFUND study. Eur J Pain 2010;
55. Rogers MW, Wilder FV. The association of BMI and knee pain
among persons with radiographic knee osteoarthritis: a cross-
sectional study. BMC Musculoskelet Disord 2008;9:163.
56. Somers TJ, Keefe FJ, Pells JJ, Dixon KE, Waters SJ, Riordan
PA, et al. Pain catastrophizing and pain-related fear in osteo-
arthritis patients: relationships to pain and disability. J Pain
Symptom Manage 2009;37:863–72.
57. Marks R. Obesity profiles with knee osteoarthritis: correlation
with pain, disability, disease progression. Obesity (Silver
58. Messier SP, Loeser RF, Miller GD, Morgan TM, Rejeski WJ,
Sevick MA, et al. Exercise and dietary weight loss in over-
weight and obese older adults with knee osteoarthritis: the
Arthritis, Diet, and Activity Promotion Trial. Arthritis Rheum
59. Somers TJ, Keefe FJ, Carson JW, Pells JJ, Lacaille L. Pain
catastrophizing in borderline morbidly obese and morbidly
obese individuals with osteoarthritic knee pain. Pain Res
60. Pierce JP, Natarajan L, Caan BJ, Parker BA, Greenberg ER, Flatt
SW, et al. Influence of a diet very high in vegetables, fruit, and
fiber and low in fat on prognosis following treatment for
breast cancer: the Women’s Healthy Eating and Living
(WHEL) randomized trial. JAMA 2007;298:289–98.
61. Beresford SA, Johnson KC, Ritenbaugh C, Lasser NL,
Snetselaar LG, Black HR, et al. Low-fat dietary pattern and
risk of colorectal cancer: the Women’s Health Initiative Ran-
domized Controlled Dietary Modification Trial. JAMA 2006;
Modifiable Risk Factors for Physical Inactivity in Knee OA 61