Osteoarthritis and Cartilage 19 (2011) 829e839

Occupational risk factors for osteoarthritis of the knee: a meta-analysis

D.F. McWilliams y *, B.F. Leeb z, S.G. Muthuri y, M. Doherty y, W. Zhang y
y Academic Rheumatology, University of Nottingham, UK
z 2nd Department of Medicine, State Hospital Stockerau, Centre for Rheumatology, Lower Austria & Karl Landsteiner Institute for Clinical Rheumatology,
A-2000 Stockerau, Landstrasse 18, Austria

a r t i c l e i n f o s u m m a r y

Article history: Introduction: Systematic reviews agree that knee osteoarthritis (OA) is related to occupational activities,
Received 16 November 2010 but have not quantified the overall risks.
Accepted 25 February 2011 Methods: Systematic review of observational studies of knee OA and occupation. Job titles, elite sport, heavy
work, kneeling, and other activities were included. Relative risk estimate and 95% confidence interval (CI)
Keywords: compared to sedentary work were retrieved or calculated for meta-analysis. Publication bias was examined
Meta-analysis
with Egger tests and heterogeneity was determined with I2 values and Q tests. Subgroup analysis was per-
Osteoarthritis
formed to examine causes of heterogeneity. A random effects model was performed to combine the data.
Knee
Occupation
Results: Studies of knee OA (n ¼ 51), persistent knee pain (n ¼ 12) and knee OA progression (n ¼ 3) were
Pain retrieved. Occupational risks for knee OA were examined in a total of 526,343 subjects in 8 cohort/
prospective/longitudinal studies, 25 cross-sectional studies and 18 case control studies. The overall odds
ratio (OR) was 1.61 (95% CI 1.45e1.78) with significant heterogeneity (I2 ¼ 83.6%). Study designs showed
a positive association between knee OA and occupational activities; cohort (OR 1.38, 95% CI 1.10e1.74),
cross-sectional (OR 1.57, 95% CI 1.37e1.81) and case control (OR 1.80, 95% CI 1.48e2.19). Overall there was
evidence of publication bias (P < 0.0001) which was apparent in the cross-sectional and case control
studies (P < 0.0001 and P ¼ 0.0247 respectively).
Conclusions: Some occupational activities increase the risk of knee OA, although the influences of
publication bias and heterogeneity are important limitations of this study. Prospective studies would
greatly improve the evidence base.
Ó 2011 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.

Introduction Occupational activities that are consistently associated with knee OA

Knee pain and knee osteoarthritis (OA) are common chronic
problems, leading to loss of work, disability and joint surgery. As no
simple, reliable therapies are currently available for these problems
it is appropriate to look at risk factors that are potentially modifi-
able. Recent systematic reviews all conclude that evidence of an
association between knee OA and occupational activities exists1e5.
However, we are not aware of any published quantitative synthesis
of the research which could place occupational activities in the
context of other known risk factors.
The mechanisms linking occupation to knee OA are believed to be
biomechanical. Forces across the knee can be measured in many
activities found in strenuous jobs6,7. Squatting and knee flexion forces
have also been measured in dairy farm workers8. Also, increased
forces across the knee joint related to knee mal-alignment and foot
angulation have been studied in relation to knee OA progression9e11.
* Address correspondence and reprint requests to: D.F. McWilliams, Academic
Rheumatology, Clinical Sciences Building, City Hospital, Nottingham NG5 1PB, UK.
Tel: 44-115-8231942; Fax: 44-115-8231757.
E-mail address: dan.mcwilliams@nottingham.ac.uk (D.F. McWilliams).
in the literature are kneeling, squatting, lifting/carrying and heavy
standing work12, which are often measured within the same
study13,14, and compared to those with sedentary employment or
working at low levels of activity. Many specific occupations have been
linked to knee OA, such as floor layers15, miners5,16, dockers17 and
sports at the elite level18. Although a wide variety of different occu-
pations and activities have been investigated, the proposed mecha-
nisms for many include biomechanical forces across the knee joint.
Therefore it may be appropriate to synthesise the relevant research to
see if the working population would benefit from a reduction in such
risks. This meta-analysis included observational epidemiology
studies of adult men and women when occupational activities and
risk of knee OA were compared.
Methods
Literature search
Systematic literature searches were performed using MEDLINE
(1950e), Web of Science (1945-), EMBASE (1980-) and PubMed

1063-4584/$ e see front matter Ó 2011 Osteoarthritis Research Society International. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.joca.2011.02.016
830 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839
(1865-). The systematic search strategy involved combining searches
for knee OA or knee pain with searches for occupational exposures
and epidemiological study designs. Search terms related to our
chosen words and exploded terms from MEDLINE were included. The
MEDLINE search was optimised to maximise coverage by comparison
with the reference lists of recent systematic reviews1,4. This search
was then used as the template for all of the other search engines,
where it was replicated as closely as possible (see Appendix 2).
Google and/or PubMed searches were performed in an ad hoc
manner for prominent authors in the field, well-known cohort
studies of OA, common keywords, and other studies online. National
Institute for Occupational Safety and Health Technical Information
Center -2 database (NIOSHTIC-2) was searched for “knee osteoar-
thritis”. Additional reports were examined from reference lists of
articles and the automated suggestions of related citations from
search engine results.
All languages were accepted by the search strategy, but the
literature search only contained English language terms.
Reports were downloaded into Reference Manager version 10.
The time period covered by the systematic literature search was
1950 to 1st July 2010.
Study selection and inclusion/exclusion criteria
Each article title was read and screened for potential relevance to
the systematic review. Those that appeared relevant had abstracts
retrieved, which were read. If the abstract appeared relevant, then
full text was retrieved of each report of interest.
To be included, full text reports had to be epidemiological studies
of occupational risks for knee OA or knee pain. Studies on the risk of
meniscal tears were excluded as being more likely to be related to
injury than OA. These included studies of heavy/arduous work,
specific job titles thought to be at increased risk, elite sport, kneeling,
squatting, climbing stairs, lifting/carrying and work while standing.
Some studies grouped activities together, such as kneeling/squatting,
and these were included in the knee strain subgroup if separate risks
were not reported. Structural and symptomatic knee OA were
acceptable measures of disease, including classification through
interviews and questionnaires known to be valid for detecting knee
OA. Symptomatic knee OA included any diagnosis (with or without
radiographic verification) that took knee pain into account, such as
knee arthroplasty due to OA; diagnosis by questionnaire about
symptoms; diagnosis by a general practitioner; or after a clinical
examination. Asymptomatic knee OA was taken when diagnosis was
by radiographic score alone, and clinical symptoms/pain were not
reported or studied. When different measures of knee OA were
reported, we chose symptomatic over asymptomatic knee OA. In
some studies, symptomatic knee OA data was used for calculating
the overall risk for meta-analysis, but the asymptomatic knee OA
data was used for the asymptomatic subgroup analysis. Studies with
adjusted or unadjusted risk estimates, and an appropriate measure
of spread, were included, as were studies where the authors
provided enough data for us to calculate unadjusted risks and 95%
confidence intervals (CI). Relative risks (RR) were calculated for
cohort and cross-sectional studies and OR for case control studies.
Reports were excluded if they examined leisure-time activities, non-
elite sports, exercise, regular worshipping activities involving
kneeling, housework or work performed while sitting (including
driving). Knee pain research in groups with a mean age less than 40
or where persistent pain was not clearly reported was also excluded.
In reports where the prevalence of knee problems was split into age
strata, we calculated risks only for those aged 40 years or more. One
report was chosen per study to obtain a relative risk for the meta-
analysis. When more than one report was eligible, the choice was
based upon study design, occupation being the primary research
target, adjusted statistical analysis (the relative risk adjusted for the
most confounders was preferred) and other factors maximising
credibility. One pooled relative risk from this report was put into the
main meta-analysis to represent this study, and it was calculated
from all of the eligible activities/jobs and levels of exposure in that
report. The order of preference for RR went from knee OA, to knee
pain, and then to knee OA progression (which maximised the
number of knee OA studies, but slightly reduced other categories).
RR for subgroup analysis were calculated once for each study, using
data from one report chosen in a similar manner.
Data extraction
English language data were extracted and coded by one investi-
gator (DMcW) and a random sample of 10% of the included studies
was validated by another (SM). Chinese articles (WZ) and German
articles (BFL) had data extracted by other investigators and we sought
data extraction in other languages from other rheumatology
researchers.
Study characteristics were recorded and are presented in
Appendix 1. Characteristics recorded included study design, setting
from which the control group was selected, size of groups, whether
occupational risks were the primary outcome, mean age, gender
split, outcome measured, funding source and country of study.
Ethnicity was not generally reported in the retrieved manuscripts,
and so ethnic differences in occupational risk could not be explicitly
compared. The outcome was defined as symptomatic knee OA,
asymptomatic knee OA or knee pain.
Quality assessment
Quality assessment was performed by sub-group analysis, as per
guidelines for publishing Meta-analysis Of Observational Studies in
Epidemiology (MOOSE)19. The setting for which the control group
was selected was used to define the setting of case control studies.
Data analysis
Each study had a pooled risk calculated using random effects for
all occupational exposures on the inclusion criteria and for all levels
of exposure compared to the reference. The meta-analysis of occu-
pational risks for knee OA is presented using odds ratios (ORs) and
95% CI’s on a forest plot, and the heterogeneity was examined using
the Q test and I2 statistic. A funnel plot was used to investigate
publication bias and an Egger test was performed. Statistical signifi-
cance was conferred by P < 0.05. For subgroup analyses, the pooled
risk for each activity or exposure was calculated per study using
random effects. As many studies measured more than one activity,
the total of risks in subgroup analyses exceeds the total studies. Meta-
regression was performed using the methodology of Lipsey and
Wilson20 using macros designed for SPSS v14 (SPSS Inc). All other
analyses were performed using StatsDirect v.2.7.8 (StatsDirect Ltd).
Results
Our literature search identified 51 studies of knee OA (sum-
marised in Appendix 1), 12 studies of knee pain21e32 and 3 studies
related to knee OA progression33e35 that met our inclusion criteria.
A flow diagram shows the study selection process (Fig. 1) and the
characteristics of the selected knee OA studies are tabulated
(Table I). The included knee OA studies were classified by design as
cohort/prospective/longitudinal (n ¼ 8, 478,408 people), cross-
sectional (n ¼ 25, 34,285 people) and case control (n ¼ 18, 13,650
people). A total of 18 studies reported more than one type of
occupational exposure, often a mixture of kneeling, lifting, carrying,
 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839 831

Fig. 1. Flow diagram of study selection.

squatting or other knee-bending activities. Risks derived from studies (bias ¼ 2.23, P < 0.0001) and in case control studies
these multiple exposures were pooled to give one overall risk per (bias ¼ 2.75, P ¼ 0.0247), but not in cohort studies (bias ¼ 1.53,
study for the main meta-analysis. The others consisted of reports P ¼ 0.4355). There was also a significant heterogeneity across all
assessing risks exclusively from a particular job title (n ¼ 10), elite studies (I2 ¼ 83.6%, P < 0.0001). The synthesis of all studies of occu-
sport (n ¼ 11), heavy/arduous/manual work (n ¼ 8) and knee- pational risks for knee OA yielded an OR of 1.61 (95% CI 1.45e1.78),
straining activities (n ¼ 4). The mean ages of study participants and a forest plot of the studies is presented in Fig. 3.
ranged from 38 to 79 in those studies where data was reported and Subgroup analyses are summarised in Table II. We observed that
the mean was 60.2 years. Many studies sampled men or women case control studies gave higher estimated risks than cross-sectional
exclusively, due to their different employment patterns, but the and cohort studies, and the hospital-based studies gave higher risk
mean percentage of women overall was 39.4%. estimates than the community-based studies. Subgroup analysis of
Figure 2 shows the funnel plot revealing a pattern consistent with standing work suggested that it was not a significant risk for knee OA,
publication bias for occupation and knee OA studies. The Egger test whereas the other subgroups associated with a positive risk. Studies
performed for all studies was statistically significant (bias ¼ 2.39, measuring symptomatic knee OA gave similar risks to those that
P < 0.0001), and so each study design was analysed on its own. A measured structural OA only, and reporting occupation as a primarily
statistically significant bias statistic was reached in cross-sectional studied risk factor did not alter the results. The studies with unadjusted
analyses gave slightly higher risk estimates than those adjusted for
Table I confounders. The risk of knee OA only differed slightly between
Occupation and knee OA study characteristics current job, lifetime work, longest job and early adulthood. When all
Cohort Cross- Case control All studies male and female data were analysed separately, the risks were similar.
sectional Several studies of occupational activities examined a possible
No. studies 8 25 18 51 graded response relationship with the risk of knee OA. However,
No. subjects 478,408 34,285 13,650 526,343 exposure was measured in different ways by different studies, which
Mean age (range) 49 (38e68) 59 (47e79) 66 (52e75) 60 (38e79)
did not allow more than 3e4 similar studies to be compared. In
Female (%) 21 34 57 39
Mean BMI N/A* 25.5 25.8 25.6 general, the levels of activity classified by the investigators appeared
Hospital based (%) 0 8 22 12 to follow a graded response with OA risk (data not shown), but formal
Knee OA measured analysis was not possible.
Symptomatic 8 10 18 36 Meta-regression analysis of covariates was performed to assess
Asymptomatic 0 15 0 15
Main occupational activity
study characteristics for influence over the risk of knee OA in this
Job title 2 6 2 10 meta-analysis. ORs were calculated for proportion of females (OR
Heavy work 2 3 3 8 0.81, 95% CI 0.65e1.01; P ¼ 0.060), cross-sectional and case control
Elite sport 2 9 0 11 study design (OR 1.17, 95% CI 0.94e1.45; P ¼ 0.138), unadjusted
Knee strain 1 1 2 4
analysis (OR 1.04, 95% CI 0.83e1.26; P ¼ 0.703), hospital-based
Multiple exposures 1 6 11 18
Occupation as study (OR 1.40, 95% CI 1.14e1.73; P ¼ 0.002) and non-primary
Primary exposure 7 20 15 42 objective (OR 1.24, 95% CI 0.98e1.53; P ¼ 0.057).
Secondary exposure 1 5 3 9 The studies examining persistent knee pain and occupational
* Body mass index (BMI) could only be derived in one study from the cohort activities were separately meta-analysed briefly. These studies
group. were those that did not examine the risks of knee OA, and so
832 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839
Fig. 2. Funnel plot of occupational risks for knee OA. Funnel plot showing all studies of
knee OA and occupation. Asymmetry suggests publication bias. Egger test for bias ¼ 2.39,
P < 0.0001.
studied a separate population. The occupational risks for persistent
knee pain (OR 1.46, 95% CI 1.15e1.85) did not appear to be subject to
publication bias (Egger test, bias ¼ 1.29, P ¼ 0.535) but heteroge-
neity was high (I2 ¼ 90.4%, P < 0.0001) (a forest plot of this subset is
presented in Appendix 3). The knee OA progression studies were
too few to be analysed further but we calculated pooled risk esti-
mates for progression of 1.42 (95% CI, 1.09e1.85)33; 1.16 (95% CI,
0.89e1.50)35; and 1.52 (95% CI, 0.85e2.75)34.
Discussion
Occupational activities have been reported as risk factors for knee
OA in previous systematic reviews1e5. However, few of them are
quantitative, possibly due to the variety of risk factors and
measurements in the literature. This meta-analysis included
51 studies with more than 500,000 people and examined a variety of
occupational factors that appear to operate through biomechanical
pathways and increase the risk of knee OA by approximately 60%.
The smaller group of studies on persistent knee pain also revealed
a very similar risk derived from occupational exposures.
There appears to be a strong likelihood of publication bias within
the literature for occupation and knee OA. This is especially true in the
cross-sectional and case control studies where the average study size
is small (Table I). For example, for case controls, the average study size
is 758 per study (18 studies and 13,650 subjects) which tends to
produce a larger relative risk that is more likely to be published
(Fig. 2). In contrast, 8 cohort studies with 478,408 subjects e almost
90% of the total population with an average study size of almost
60,000 is less likely to produce the small study effect36, a possible
reason for no detectable publication bias. Also, case control studies
present risk as an OR, which may overestimate the relative risk of
common diseases37. Early adult life is thought to be important for the
development of OA, but recall of activities in the past may be biased or
inaccurate. We retrieved reports in different languages, but were
unsuccessful in obtaining data extraction from some languages. If
English language publication was related to more positive findings,
then this could bias our meta-analysis.
In the meta-analysis of all occupational risks and knee OA, several
studies had a single risk estimated by pooling risks from different
activities. Despite the common biomechanical pathway towards knee
OA, each separate activity measured will increase the variation of this
study. We present each activity as part of the subgroup analysis in
Table II, and most were positive risk factors for knee OA, although
heterogeneity measures were still high. Variation also occurred in the
investigators’ choices of the parameters to be measured and the
different periods of life to study. The most-common time periods
were current job, longest-held job, early adulthood (especially
common for elite sport) and lifetime work. The differences in
measurement could contribute to variability, although the current job
is likely to be similar to the longest-held job for many subjects.
The subgroup analysis of occupational risks showed the highest
estimates in studies of particular job titles, and also in elite sports
which is a group of job titles in itself. A possible reason for this could
be that those jobs with the highest risk were more likely to be
studied and reported. Possibly, high risk job titles may have been
pulled out of larger data sets and highlighted in publications. The risk
from kneeling, measured as part of occupation was moderate.
However, floor layers (included in the job description category) have
been intensely studied and found to be at high risk of knee OA38e41.
Their main occupational exposure is believed to be kneeling, and one
study has shown a reduction in knee symptoms associated with
reduced kneeling practises42. Other activities also gave positive risks
for knee OA, with the exception of standing work. Standing during
work may be tiring, but our analysis does not indicate it as a risk of
knee OA. It should also be pointed out that riskier occupations are
often performed by men. Although the subgroup analysis showed no
differences in risk between the genders, men are probably more
exposed to occupational risks. This is reflected in the high number of
men studied in the literature. A propensity towards knee injuries in
risky occupations must not be ignored. One study found that
adjusting for knee injuries explained the sports-related risks for knee
OA43, although others included in this analysis found that adjust-
ment did not explain all of the association with occupation.
Hospital-based studies yielded higher risks of knee OA than
community-based studies, although the number of hospital-based
studies was small. The increased risk estimates may be due to more
severe disease or related to controls choosing sedentary work. Where
occupation was not judged to be a primary outcome, the risks of knee
OA were slightly higher, which might have occurred if secondary
results were reported due to being statistically significant. Our search
strategy was English-language only, meaning that studies will have
been missed. We were unable to examine any ethnic differences in
occupational risk. The region where the study was performed could
be taken to represent to primary ethnicity of the population at risk,
although almost no studies explicitly stated ethnic background
(making such interpretations imperfect). We may have also missed
some occupational risks from our meta-analysis. We focussed on
those where a clear biomechanical pathway could be hypothesised
due to load-bearing, pressure, knee strain or forces across the knee
joint. Vibration from machinery, working in uncomfortable positions,
driving, sitting activities and walking were not included. Leisure-time
activities, such as exercise, were not included in this meta-analysis, as
we hypothesised that they were more likely to be beneficial than
harmful. The occupational activities we included were those that
subjects would have probably been obliged to perform, and less likely
to consistently avoid.
The literature studying occupation and knee OA could be
strengthened by more cohort or longitudinal/prospective studies.
Many occupational activities are measured in detail in the case
control or cross-sectional studies, but their effects on incident knee
OA have not been adequately investigated. These types of studies
would improve the current evidence base. It may also be of interest
to assess the interaction between occupation and known risk factors
for knee OA. Some studies have looked at body mass index (BMI)
interacting with occupation13,44, and we have found an interaction
between knee mal-alignment and occupational risks (but not BMI) in
the Genetics Osteoarthritis and Lifestyle (GOAL) case control study9.
Both may influence the forces acting across the knee joint, which
may affect the development of knee OA.
 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839 833

Fig. 3. Forest plot showing all occupation and knee OA studies grouped by design with sub-totals shown for each type (cohort, cross-sectional and case control), and arranged
in chronological order within each group.
834 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839

Table II
Subgroup analyses for occupation and knee OA

Subgroups No. subjects Heterogeneity % (Cochran Q, P value) Summary meta-analysis

Study design
Cohort 478,408 81.3 (<0.0001) 1.38 (1.10e1.74)
Cross-sectional 34,285 77.7 (<0.0001) 1.57 (1.37e1.81)
Case control 13,650 87.8 (<0.0001) 1.80 (1.48e2.19)
Setting
Community 521,132 81.6 (<0.0001) 1.52 (1.38e1.69)
Hospital 5,211 90.6 (<0.0001) 2.65 (1.62e4.36)
Occupation measured as
Primary exposure 505,918 81.3 (<0.0001) 1.60 (1.44e1.77)
Secondary exposure 20,425 88.9 (<0.0001) 1.74 (1.28e2.36)
Occupational activities
Job title 233,469 77.1 (<0.0001) 1.99 (1.63e2.43)
Heavy or manual work 546,853 80.9 (<0.0001) 1.45 (1.20e1.76)
Elite sports 9,703 55.9 (0.08579) 1.72 (1.35e2.20)
Kneeling 9,236 68.2 (0.0009) 1.30 (1.03e1.63)
Squatting 13,181 25.5 (0.2013) 1.40 (1.21e1.61)
Lifting/carrying 11,833 77.7 (<0.0001) 1.58 (1.28e1.94)
Climbing stairs 16,253 72.5 (<0.0001) 1.29 (1.08e1.55)
Standing work 7,896 80.8 (<0.0001) 1.11 (0.81e1.51)
Knee bending/straining 262,073 72.1 (0.0007) 1.60 (1.15e2.21)
OA definition criteria
Symptomatic 510,487 86.9 (<0.0001) 1.62 (1.43e1.83)
Asymptomatic 19,485 77.4 (<0.0001) 1.57 (1.33e1.86)
Confounding variables
Adjusted analysis 514,887 81.0 (<0.0001) 1.51 (1.36e1.67)
Unadjusted analysis 11,073 88.4 (<0.0001) 1.94 (1.42e2.67)
Gender
Male data 446,057 69.5 (<0.0001) 1.53 (1.36e1.73)
Female data 63,666 52.2 (0.0052) 1.61 (1.42e1.82)
Region
Europe 506,525 83.6 (<0.0001) 1.71 (1.48e1.96)
North America 12,747 74.2 (0.0038) 1.32 (1.09e1.60)
Oriental 5,631 84.5 (<0.0001) 1.63 (1.20e2.23)
Rest of World 1,360 88.6 (<0.0001) 1.40 (0.83e2.35)
Funding
Charity/Government/Foundation 485,520 82.4 (<0.0001) 1.52 (1.34e1.73)
Trade union/Industry 19,759 68.0 (0.0139) 1.48 (1.10e1.98)
Not reported 20,723 89.8 (<0.0001) 1.82 (1.40e2.35)
Occupational exposure time period
Current job 36,594 87.1 (<0.0001) 1.83 (1.42e2.35)
Longest-held job 8,177 63.8 (0.0073) 1.60 (1.25e2.04)
Early adulthood 11,777 59.6 (0.008) 1.73 (1.34e2.22)
Lifetime 14,014 89.8 (<0.0001) 1.47 (1.24e1.74)
Knee compartment
Tibiofemoral 18,319 78.4 (<0.0001) 1.41 (1.17e1.71)
Tibiofemoral or patellofemoral 12,052 78.8 (<0.0001) 1.47 (1.12e1.94)
Not reported 495,589 84.8 (<0.0001) 1.76 (1.52e2.03)

Subgroup analysis to show if heterogeneity is related to certain study characteristics. The percentage values for heterogeneity are from the I2 test and the P values for are from
the Cochran Q test. Data from some studies45e47 included in the asymptomatic knee OA subgroup were not included in the overall meta-analysis, because the symptomatic
knee OA risks were included instead. Oriental countries (n ¼ 9) included China, Japan, Thailand, Hong Kong and Korea. The Rest of the World (n ¼ 4) consisted of Australia, Iran,
Tunisia and Morocco only.

Our search strategy only returned a small number of OA Data extraction and validation: Zhang, Leeb, McWilliams, Muthuri
progression studies. Although it is possible that our strategy was sub- Interpretation of the data and analyses: Zhang, Leeb, Doherty,
optimal for this type of research, it does appear that future studies McWilliams, Muthuri
should consider OA progression. Studies were preferentially included Drafting and revision of the manuscript: Zhang, Leeb, Doherty,
into the knee OA group where possible, reducing both knee pain and McWilliams, Muthuri
OA progression groups. When we examined knee pain, we excluded Approval of the final manuscript: Zhang, Leeb, Doherty,
a lot of studies where the duration of pain was either too short to McWilliams, Muthuri
define as persistent or not clearly reported. Our brief meta-analysis of
Conflict of interest
knee pain shows a risk estimate similar to knee OA, possibly because
We declare no conflicts of interest.
persistent joint pain in the over 40’s implies OA. The knee pain
studies we analysed were also very heterogeneous, and so must be
interpreted cautiously. Acknowledgements

Author contributions We would like to acknowledge Professor Stefan Lohmander, and

Professor Tiraje Tuncer for translation and data extraction; and
Conception, design, inclusion/exclusion criteria of studies: Joanna Ramowski and Helen Richardson for additional support. The
Zhang, Leeb, Doherty, McWilliams, Muthuri authors report no external financial support for this work (and no
Statistical analysis: Zhang, McWilliams other agencies approved the final version of the manuscript).
 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839 835

Appendix 1

Summary of studies included in knee OA and occupation meta-analysis

Author & year Country Design Setting Subjects Mean Female Mean Primary? Knee Category Risk (95% CI) Adj?
age (%) BMI OA diag
Felson 199147 USA c comm 1,366 e 59 e y symp Multiple 0.94 (0.56e1.58) adj
Vingard 199148 Sweden c comm 250,217 e 21 e y symp Knee strain 1.4 (1.16e1.69) adj
Sarna 199749 Finland c comm 4,325 e 0 e y symp Elite sport 1.83 (1.14e2.93) un
Kujala 199950 Finland c comm 457 49 0 23.6 y symp Elite sport 1.79 (1.1e3.54) adj
Jarvholm 200841 Sweden c comm 204,741 e 0 e y symp Job title 1.91 (1.57e2.33) adj
Kaerlev 200851 Denmark c comm 14,801 38 0 e y symp Job title 1.16 (0.95e1.42) adj
Verweij 200952 Netherlands c comm 1,678 68 32 e n asym Heavy work 0.95 (0.81e1.11) un
Toivanen 201053 Finland c comm 823 42 55 e y symp Heavy work 1.91 (0.97e3.76) adj
Kohatsu 199054 USA cc comm 94 71 39 25.4 y symp Heavy work 2.75 (1.43e5.26) un
Elsner 199655 Germany cc comm 383 e 43 e y symp Multiple 1.64 (1.31e2.05) un
Sahlstrom 199756 Sweden cc comm 729 75 e e y symp Knee strain 1.1 (0.7e1.8) adj
Sandmark 200057 Sweden cc comm 1,173 e 48 e y symp Multiple 1.77 (1.54e2.03) adj
Lau 200058 Hong Kong cc comm 1,536 e 1 e y symp Multiple 2.43 (1.47e4.03) adj
Coggon 200013 UK cc comm 1,036 72 60 e y symp Multiple 1.68 (1.42e1.99) adj
Manninen 200259 Finland cc comm 805 67 76 25.0 y symp Multiple 1.15 (0.98e1.36) adj
Dawson 200360 UK cc comm 111 e 100 e n symp Multiple 3.06 (1.87e4.99) un
Yoshimura 200461 Japan cc comm 202 73 100 23.2 y symp Multiple 1.25 (0.93e1.68) adj
Holmberg 200462 Sweden cc comm 1,473 63 57 e y symp Job title 1.40 (1.09e1.79) adj
Cebi 200563 Turkey cc hos 100 60 96 e y symp Multiple 8.24 (4.46e15.25) un
Yoshimura 200664 Japan cc comm 74 70 0 23.3 y symp Heavy work 6.2 (1.4e27.5) adj
Riyazi 200865 Netherlands cc comm 727 59 73 26.0 n symp Heavy work 4.3 (1.7e11.4) adj
Mounach 200866 Morocco cc comm 190 60 73 29.1 y symp Multiple 0.67 (0.48e0.94) un
Dahaghin 200967 Iran cc comm 970 52 67 28.6 y symp Multiple 1.07 (0.93e1.22) adj
He 200968 China cc hos 247 e 30 e n symp Knee strain 18.09 (5.86e55.84) adj
Franklin 201069 Iceland cc hos 2,490 72 57 26.2 y symp Job title 1.79 (1.16e2.78) adj
Klussmann 201070 Germany cc hos 1,310 59 57 e y symp Multiple 1.69 (1.35e2.12) adj
Lawrence 195516 UK cs comm 362 e 0 e y asym Multiple 2.17 (1.05e4.47) un
Kellgren 195871 UK cs comm 380 e 54 e y asym Job title 2.09 (1.28e3.42) un
Partridge 196817 UK cs comm 377 e 0 e y symp Job title 2.24 (1.1e4.56) un
Klunder 198072 Denmark cs comm 114 56 0 e y asym Elite sport 1.58 (1.03e2.43) un
Lindberg 198773 Sweden cs comm 1,122 65 0 e y asym Job title 2.76 (1.04e7.31) un
Anderson 198874 USA cs comm 5,193 e 53 e n asym Multiple 1.62 (1.2e2.18) adj
Konradsen 199075 Denmark cs comm 54 58 0 e y asym Elite sport 1.30 (0.74e2.28) un
Bagge 199176 Sweden cs comm 340 79 61 e y asym Elite sport 0.89 (0.69e1.14) un
Roos 199477 Sweden cs comm 643 56 0 e y asym Elite sport 5.5 (1.33e22.75) un
Vingard 199578 Sweden cs comm 469 100 y symp Elite sport 2.8 (0.7e11.0) adj
Spector 199679 UK cs comm 1,058 54 100 25.3 y asym Elite sport 2.48 (1.33e4.63) adj
Deacon 199746 Australia cs comm 100 55 0 26.1 y symp Elite sport 4.0 (1.9e8.2) adj
Sandmark 200080 Sweden cs comm 1,083 57 50 e y symp Job title 2.99 (2.02e4.42) un
Kettunen 200181 Finland cs comm 1,375 64 0 e y symp Elite sport 1.28 (0.95e1.73) adj
Jensen 200582 Denmark cs comm 966 48 0 25.5 y symp Knee strain 3.90 (1.49e10.17) adj
Andrianakos 200683 Greece cs comm 8,740 47 n symp Heavy work 0.91 (0.83e1.11) adj
Tangtrakulwanich 200684 Thailand cs comm 261 60 0 23.2 y asym Multiple 1.13 (0.75e1.7) adj
Elleuch 200885 Tunisia cs hos 100 49 0 24.8 y asym Elite sport 1.93 (1.22e3.06) un
Fernandez-Lopez 200886 Spain cs comm 2,192 e 54 e n symp Heavy work 1.55 (0.92e2.60) adj
Muraki 200987 Japan cs comm 1,471 68 64 23.1 y asym Multiple 1.53 (1.25e1.88) adj
Rytter 200940 Denmark cs comm 254 53 0 26.4 y asym Job title 2.49 (1.04e5.93) adj
Wang 200988 China cs hos 964 55 62 24.6 n symp Job title 1.22 (1.0e1.48) adj
Allen 201045 USA cs comm 2,546 64 66 30.7 y symp Multiple 1.12 (1.04e1.20) adj
Kim 201089 Korea cs comm 573 70 54 24.6 n asym Heavy work 2.14 (1.19e3.84) adj
Bernard 201090 USA cs comm 3,548 62 69 26.6 y asym Multiple 1.30 (1.13e1.49) adj

c ¼ cohort, cs ¼ cross-sectional, cc ¼ case control, comm ¼ community-based, hos ¼ hospital-based, symp ¼ symptomatic knee OA diagnostic criteria, asym ¼ asymptomatic
knee OA diagnostic criteria, adj ¼ adjusted analysis, un ¼ unadjusted analysis, y ¼ yes, n ¼ no.
836 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839

Appendix 2 62. 55 or 56 or 57 or 58 or 59 or 60 or 61
63. work-related.mp.
MEDLINE search strategy 64. physical demands.mp.
65. exp Workplace/or working condition$.mp.
1. knee osteoarthritis.mp. or exp Osteoarthritis, Knee/ 66. activit$.mp.
2. knee osteoarthrosis.mp. 67. activities.mp. or exp "Activities of Daily Living"/or exp Leisure
3. gonarthrosis.mp. Activities/
4. gonarthritis.mp. 68. exp "Moving and Lifting Patients"/or lifting.mp. or exp Lifting/
5. knee pain.mp. or exp Weight Lifting/
6. 1 or 2 or 3 or 4 or 5 69. squatting.mp.
7. osteoarthritis.mp. or exp Osteoarthritis/ 70. crawling.mp.
8. osteoarthrosis.mp. 71. kneeling.mp.
9. arthrosis.mp. 72. kneel$.mp.
10. osteophyte.mp. or exp Osteophyte/ 73. squat$.mp.
11. joint space narrowing.mp. 74. crawl.mp.
12. degenerative joint disease$.mp. 75. crawl$.mp.
13. 7 or 8 or 9 or 10 or 11 or 12 76. carrying.mp.
14. musculoskeletal.mp. or exp Musculoskeletal Diseases/ 77. climbing.mp.
15. pain.mp. or exp Pain/ 78. climb$.mp.
16. symptom$.mp. 79. exp Walking/or walking.mp.
17. disorder.mp. 80. stairs.mp.
18. 15 or 16 or 17 81. stair$.mp.
19. 14 and 18 82. standing.mp.
20. knee.mp. or exp Knee/ 83. work pace.mp.
21. 19 and 20 84. physical exertion.mp. or exp Physical Exertion/
22. 13 and 20 85. forceful work.mp.
23. 6 or 21 or 22 86. forceful exertion.mp.
24. cohort studies.mp. or exp Cohort Studies/ 87. tasks.mp.
25. cohort stud$.mp. 88. cumulative strain.mp.
26. prospective stud$.mp. 89. overuse$.mp.
27. relative risk$.mp. 90. bending.mp.
28. incidence.mp. or exp Incidence/ 91. bend$.mp.
29. incident.mp. 92. manual lifting.mp.
30. 24 or 25 or 26 or 27 or 28 93. heavy lifting.mp.
31. exp Epidemiology/or epidemiology.mp. 94. heavy lift$.mp.
32. 29 and 31 95. shift work.mp.
33. 30 or 32 96. exp Vibration/or vibration.mp.
34. case-control studies.mp. or exp Case-Control Studies/ 97. physical work.mp. or exp Workload/
35. case control stud$.mp. 98. work load.mp.
36. retrospective stud$.mp. 99. gardening.mp. or exp Gardening/
37. exp Odds Ratio/ 100. 63 or 64 or 65 or 66 or 67 or 68 or 69 or 70 or 71 or 72 or 73 or 74
38. odds ratio$.mp. or 75 or 76 or 77 or 78 or 79 or 80 or 81 or 82 or 83 or 84 or 85 or
39. odds.mp. 86 or 87 or 88 or 89 or 90 or 91 or 92 or 93 or 94 or 95 or 96 or 97
40. community control$.mp. or 98 or 99
41. referent$.mp. 101. exp Mining/or miner.mp. or exp Coal Mining/
42. 34 or 35 or 36 or 37 or 38 or 39 or 40 or 41 102. football$.mp. or exp Football/
43. cross-sectional studies.mp. or exp Cross-Sectional Studies/ 103. soccer.mp. or exp Soccer/
44. cross sectional stud$.mp. 104. exp Running/or runner.mp.
45. risk.mp. or exp Risk/ 105. sport$.mp. or exp Sports/
46. prevalence.mp. or exp Prevalence/ 106. athlet$.mp.
47. 43 or 44 or 45 or 46 107. construction worker.mp.
48. questionnaire.mp. or exp Questionnaires/ 108. construction work$.mp.
49. 31 and 48 109. gardener.mp.
50. 47 or 49 110. floor layer.mp.
51. risk factor.mp. or exp Risk Factors/ 111. floor lay$.mp.
52. 48 and 51 112. carpet fitter.mp.
53. 50 or 52 113. carpet lay$.mp.
54. 33 or 42 or 53
114. prayer.mp.
55. occupation$.mp. or exp Occupations/
115. praying.mp.
56. employee$.mp.
116. 101 or 102 or 103 or 104 or 105 or 106 or 107 or 108 or 109 or
57. exp Employment/or employment.mp.
110 or 111 or 112 or 113 or 114 or 115
58. industr$.mp.
117. 62 or 100 or 116
59. work.mp. or exp Work/
60. workers.mp. 118. 33 or 42 or 54
61. job.mp. 119. 23 and 117 and 118
 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839
Appendix 3
Forest plot of studies of occupational risks for persistent knee pain in those over 40 years old. Knee OA studies were not included.
References
1. Blagojevic M, Jinks C, Jeffery A, Jordan KP. Risk factors for onset
of osteoarthritis of the knee in older adults: a systematic
review and meta-analysis. Osteoarthritis Cartilage 2010;18(1):
24e33.
2. Vignon E, Valat JP, Rossignol M, Avouac B, Rozenberg S,
Thoumie P, et al. Osteoarthritis of the knee and hip and activity:
a systematic international review and synthesis (OASIS). Joint
Bone Spine 2006;73(4):442e55.
3. Lequesne M. Sport practice and osteoarthritis of the limbs.
Sci Sports 2004;19(6):281e5.
4. Jensen LK, Eenberg W. Occupation as a risk factor for knee
disorders. Scand J Work Environ Health 1996;22(3):165e75.
5. McMillan G, Nichols L. Osteoarthritis and meniscus disorders
of the knee as occupational diseases of miners. Occup Environ
Med 2005;62(8):567e75.
6. Faber GS, Kingma I, van Dieen JH. Bottom-up estimation of
joint moments during manual lifting using orientation sensors
instead of position sensors. J Biomech 2010;43(7):1432e6.
7. Gullett JC, Tillman MD, Gutierrez GM, Chow JW.
A biomechanical comparison of back and front squats in
healthy trained individuals. J Strength Cond Res 2009;23(1):
284e92.
8. Nonnenmann MW, Anton DC, Gerr F, Yack HJ. Dairy farm
worker exposure to awkward knee posture during milking and
feeding tasks. J Occup Environ Hyg 2010;7(8):483e9.
9. McWilliams DF, Doherty S, Maciewicz RA, Muir KR, Zhang WY,
Doherty M. Self-reported knee and foot alignments in early
adult life and risk of osteoarthritis. Arthritis Care Res
2010;62(4):489e95.
10. Chang A, Hayes K, Dunlop D, Hurwitz D, Song J, Cahue S, et al.
Thrust during ambulation and the progression of knee osteo-
arthritis. Arthritis Rheum 2004;50(12):3897e903.
837
11. Chang A, Hurwitz D, Dunlop D, Song J, Cahue S, Hayes K, et al.
The relationship between toe-out angle during gait and
progression of medial tibiofemoral osteoarthritis. Ann Rheum
Dis 2007;66(10):1271e5.
12. Cooper C, Coggon D. Physical activity and knee osteoarthritis.
Lancet 1999;353(9171):2177e8.
13. Coggon D, Croft P, Kellingray S, Barrett D, McLaren M,
Cooper C. Occupational physical activities and osteoarthritis of
the knee. Arthritis Rheum 2000;43(7):1443e9.
14. Cooper C, Mcalindon T, Coggon D, Egger P, Dieppe P. Occu-
pational activity and osteoarthritis of the knee. Ann Rheum Dis
1994;53(2):90e3.
15. Tanaka S, Smith AB, Halperin W, Jensen R. Carpet-layer’s knee.
N Engl J Med 1982;307(20):1276e7.
16. Lawrence JS. Rheumatism in coal miners. III. Occupational
factors. Br J Ind Med 1955;12:249e61.
17. Partridge REH, Duthie JJR. Rheumatism in dockers and civil
servants: a comparison of heavy manual and sedentary
workers. Ann Rheum Dis 1968;27:559e68.
18. Kujala UM, Kaprio J, Sarna S. Osteoarthritis of weight-bearing
joints of lower-limbs in former elite male-athletes. BMJ
1994;308(6923):231e4.
19. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D,
et al. Meta-analysis of observational studies in epidemiology:
a proposal for reporting. Meta-analysis of observational studies
in epidemiology (MOOSE) group. JAMA 2000;283(15):2008e12.
20. Lipsey MW, Wilson DB. The way in which intervention studies
have "personality" and why it is important to meta-analysis.
Eval Health Prof 2001;24(3):236e54.
21. Bergenudd H, Nilsson B, Lindgarde F. Knee pain in middle age
and its relationship to occupational work load and psychoso-
cial factors. Clin Orthop Relat Res 1989;245:210e5.
838 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839
22. Holmberg S, Stiernstrom EL, Thelin A, Svardsudd K. Musculo-
skeletal symptoms among farmers and non-farmers: a pop-
ulation-based study. Int J Occup Environ Health 2002;8(4):
339e45.
23. Lemasters GK, Atterbury MR, Booth-Jones AD, Bhattacharya A,
Ollila-Glenn N, Forrester C, et al. Prevalence of work related
musculoskeletal disorders in active union carpenters. Occup
Environ Med 1998;55(6):421e7.
24. Miranda H, Viikari-Juntura E, Martikainen R, Riihimaki H.
A prospective study on knee pain and its risk factors. Osteo-
arthritis Cartilage 2002;10(8):623e30.
25. O’Reilly SC, Muir KR, Doherty M. Occupation and knee pain:
a community study. Osteoarthritis Cartilage 2000;8(2):
78e81.
26. Raty HP, Kujala UM, Videman T, Impivaara O, Battie MC, Sarna S.
Lifetime musculoskeletal symptoms and injuries among former
elite male athletes. Int J Sports Med 1997;18(8):625e32.
27. Rytter S, Jensen LK, Bonde JP. Knee complaints and consequences
on work status; a 10-year follow-up survey among floor layers
and graphic designers. BMC Musculoskelet Disord 2007;8.
28. Sjolander P, Daerga L, Edin-Liljegren A, Jacobsson LT.
Musculoskeletal symptoms and perceived work strain among
reindeer herders in Sweden. Occup Med (Oxford) 2008;58
:572e9.
29. Sobti A, Cooper C, Inskip H, Searle S, Coggon D. Occupational
physical activity and long-term risk of musculoskeletal symp-
toms: a national survey of post office pensioners. Am J Ind Med
1997;32(1):76e83.
30. Holmstrom E, Engholm G. Musculoskeletal disorders in rela-
tion to age and occupation in Swedish construction workers.
Am J Ind Med 2003;44:377e84.
31. Thun M, Tanaka S, Smith AB, Halperin WE, Lee ST, Luggen ME,
et al. Morbidity from repetitive knee trauma in carpet and
floor layers. Br J Ind Med 1987;44(9):611e20.
32. Wickstrom G, Hanninen K, Mattsson T, Riihimaki H, Waris P,
Zitting A. Knee degeneration in concrete reinforcement
workers. Br J Ind Med 1983;40:216e9.
33. Amin S, Goggins J, Niu J, Guermazi A, Grigoryan M, Hunter DJ,
et al. Occupation-related squatting, kneeling, and heavy lifting
and the knee joint: a magnetic resonance imaging-based study
in men. J Rheumatol 2008;35(8):1645e9.
34. Schouten JS, Van den Ouweland FA, Valkenburg HA. A 12 year
follow up study in the general population on prognostic
factors of cartilage loss in osteoarthritis of the knee. Ann
Rheum Dis 1992;51(8):932e7.
35. Teichtahl AJ, Wluka AE, Wang Y, Urquhart DM, Hanna FS,
Berry PA, et al. Occupational activity is associated with knee
cartilage morphology in females. Maturitas 2010;66(1):72e6.
36. Sterne JA, Gavaghan D, Egger M. Publication and related
bias in meta-analysis: power of statistical tests and preva-
lence in the literature. J Clin Epidemiol 2000;53(11):
1119e29.
37. Modern Epidemiology. 2 edn. Philadelphia, USA: Lippincott
Williams & Wilkins 1998.
38. Kivimaki J, Riihimaki H, Hanninen K. Knee disorders in carpet
and floor layers and painters. Scand J Work Environ Health
1992;18(5):310e6.
39. Rytter S, Jensen LK, Bonde JP. Clinical knee findings in floor
layers with focus on meniscal status. BMC Musculoskelet
Disord 2008;9(144)
40. Rytter S, Egund N, Jensen LK, Bonde JP. Occupational kneeling
and radiographic tibiofemoral and patellofemoral osteoar-
thritis. J Occup Med Toxicol 2009;4:19.
41. Jarvholm B, From C, Lewold S, Malchau H, Vingard E. Inci-
dence of surgically treated osteoarthritis in the hip and knee
in male construction workers. Occup Environ Med
2008;65(4):275e8.
42. Jensen LK, Friche C. Effects of training to implement new
working methods to reduce knee strain in floor layers. A two-
year follow-up. Occup Environ Med 2008;65(1):20e7.
43. Thelin N, Holmberg S, Thelin A. Knee injuries account for the
sports-related increased risk of knee osteoarthritis. Scand J
Med Sci Sports 2006;16(5):329e33.
44. Vrezas I, Elsner G, Bolm-Audorff U, Abolmaali N, Seidler A.
Case-control study of knee osteoarthritis and lifestyle
factors considering their interaction with physical
workload. Int Arch Occup Environ Health 2010;83(3):
291e300.
45. Allen KD, Chen JC, Callahan LF, Golightly YM, Helmick CG,
Renner JB, et al. Associations of occupational tasks with knee
and hip osteoarthritis: the Johnston County osteoarthritis
project. J Rheumatol 2010;37(4):842e50.
46. Deacon A, Bennell K, Kiss ZS, Crossley K, Brukner P. Osteoar-
thritis of the knee in retired, elite Australian rules footballers.
Med J Aust 1997;166(4):187e90.
47. Felson DT, Hannan MT, Naimark A, Berkeley J, Gordon G,
Wilson PWF, et al. Occupational physical demands, knee
bending, and knee osteoarthritis e results from the Framing-
ham-study. J Rheumatol 1991;18(10):1587e92.
48. Vingard E, Alfredsson L, Goldie I, Hogstedt C. Occupation and
osteoarthrosis of the hip and knee: a register-based cohort
study. Int J Epidemiol 1991;20(4):1025e31.
49. Sarna S, Kaprio J, Kujala UM, Koskenvuo M. Health status of
former elite athletes. The Finnish experience. Aging (Milano)
1997;9(1e2):35e41.
50. Kujala UM, Sarna S, Kaprio J, Koskenvuo M, Karjalainen J. Heart
attacks and lower-limb function in master endurance athletes.
Med Sci Sports Exerc 1999;31(7):1041e6.
51. Kaerlev L, Jensen A, Nielsen PS, Olsen J, Hannerz H, Tuchsen F.
Hospital contacts for injuries and musculoskeletal diseases
among seamen and fishermen: a population-based cohort
study. BMC Musculoskelet Disord 2008;9(8).
52. Verweij LM, van Schoor NM, Deeg DJH, Dekker J, Visser M.
Physical activity and incident clinical knee osteoarthritis in
older adults. Arthritis Rheum 2009;61(2):152e7.
53. Toivanen AT, Heliovaara M, Impivaara O, Arokoski JP, Knekt P,
Lauren H, et al. Obesity, physically demanding work and trau-
matic knee injury are major risk factors for knee
osteoarthritisea population-based study with a follow-up of 22
years. Rheumatology (Oxford) 2010;49(2):308e14.
54. Kohatsu ND, Schurman DJ. Risk factors for the development of
osteoarthrosis of the knee. Clin Orthop Relat Res 1990;261:
242e6.
55. Elsner G, Nienhaus A, Beck W. Arthrosis of the knee and work--
related factors. Sozial-und Praventivmedizin 1996;41(2):
98e106.
56. Sahlstrom A, Montgomery F. Risk analysis of occupational
factors influencing the development of arthrosis of the knee.
Eur J Epidemiol 1997;13(6):675e9.
57. Sandmark H, Hogstedt C, Vingard E. Primary osteoarthrosis of
the knee in men and women as a result of lifelong physical
load from work. Scand J Work Environ Health 2000;26(1):
20e5.
58. Lau EC, Cooper C, Lam D, Chan VNH, Tsang KK, Sham A. Factors
associated with osteoarthritis of the hip and knee in Hong
Kong Chinese: obesity, joint injury, and occupational activities.
Am J Epidemiol 2000;152(9):855e62.
59. Manninen P, Heliovaara M, Riihimaki H, Suomalainen O.
Physical workload and the risk of severe knee osteoarthritis.
Scand J Work Environ Health 2002;28(1):25e32.
 D.F. McWilliams et al. / Osteoarthritis and Cartilage 19 (2011) 829e839
60. Dawson J, Juszczak E, Thorogood M, Marks SA, Dodd C,
Fitzpatrick R. An investigation of risk factors for symptomatic
osteoarthritis of the knee in women using a life course
approach. J Epidemiol Community Health 2003;57(10):823e30.
61. Yoshimura N, Nishioka S, Kinoshita H, Hori N, Nishioka T,
Ryujin M, et al. Risk factors for knee osteoarthritis in Japanese
women: heavy weight, previous joint injuries, and occupa-
tional activities. J Rheumatol 2004;31(1):157e62.
62. Holmberg S, Thelin A, Thelin N. Is there an increased risk of knee
osteoarthritis among farmers? A population-based case-control
study. Int Arch Occup Environ Health 2004;77(5):345e50.
63. Cebi E, Karkucak M, Tosun M, Cebi A, Topbas M, Capkin E. Risk
factors for hip and knee osteoarthritis. J Rheumatol Medical
Rehabil 2005;16(1):2e9.
64. Yoshimura N, Kinoshita H, Hori N, Nishioka T, Ryujin M,
Mantani Y, et al. Risk factors for knee osteoarthritis in Japa-
nese men: a case-control study. Mod Rheumatol 2006;16(1):
24e9.
65. Riyazi N, Rosendaal FR, Slagboom E, Kroon HM, Breedveld FC,
Kloppenburg M. Risk factors in familial osteoarthritis: the GARP
sibling study. Osteoarthritis Cartilage 2008;16(6):654e9.
66. Mounach A, Nouijai A, Ghozlani I, Ghazi M, Achemlal L,
Bezza A, et al. Risk factors for knee osteoarthritis in Morocco. A
case control study. Clin Rheumatol 2008;27(3):323e6.
67. Dahaghin S, Tehrani-Banihashemi SA, Faezi ST, Jamshidi AR,
Davatchi F. Squatting, sitting on the floor, or cycling: are life-
long daily activities risk factors for clinical knee osteoarthritis?
Stage III results of a community-based study. Arthritis Care Res
2009;61(10):1337e42.
68. He C-H, Tan F-Y, Xie X-M, Liu D-X, Chen Q. Risk factor inves-
tigation of knee osteoarthritis of 126 senile patients in Yuexiu
district of Guangzhou city: contrast with 121 senile patients
with no knee osteoarthritis. J Clin Rehabil Tissue Eng Res
2009;13(33):6581e4 [Chinese].
69. Franklin J, Ingvarsson T, Englund M, Lohmander S. Association
between occupation and knee and hip replacement due to
osteoarthritis: a case-control study. Arthritis Res Ther
2010;12(3):R102.
70. Klussmann A, Hansjurgen A, Nubling M, Liebers F, Quiros
Perea E, Cordier W, et al. Individual and occupational risk
factors for knee osteoarthritis: results of a case control study in
Germany. Arthritis Res Ther 2010;12:R88.
71. Kellgren JH, Lawrence JS. Osteo-arthrosis and disk degenera-
tion in an urban population. Ann Rheum Dis 1958;17:388e97.
72. Klunder KB, Rud B, Hansen J. Osteoarthritis of the hip and knee joint
in retired football players. Acta Orthop Scand 1980;51(6):925e7.
73. Lindberg H, Montgomery F. Heavy labor and the occurrence of
gonarthrosis. Clin Orthop Relat Res 1987;214:235e6.
74. Anderson JJ, Felson DT. Factors associated with osteoarthritis of hte
knee in the first national health and nutrition examination survey
(HANES I). Evidence for an association with overweight, race, and
physical demands of work. Am J Epidemiol 1988;128(1):179e89.
75. Konradsen L, Berg HEM, Sondergaard L. Long distance running
and osteoarthrosis. Am J Sports Med 1990;18(4):379e86.
839
76. Bagge E, Bjelle A, Eden S, Svanborg A. Factors associated with
radiographic osteoarthritis: results from the population study
70-year-old people in Goteborg. J Rheumatol 1991;18(8):
1218e22.
77. Roos H, Lindberg H, Gardsell P, Lohmander LS, Wingstrand H.
The prevalence of gonarthrosis and its relation to meniscec-
tomy in former soccer players. Am J Sports Med 1994;22(2):
219e22.
78. Vingard E, Sandmark H, Alfredsson L. Musculoskeletal disor-
ders in former athletes e a cohort study in 114 track and field
champions. Acta Orthop Scand 1995;66(3):289e91.
79. Spector TD, Harris PA, Hart DJ, Cicuttini FM, Nandra D,
Etherington J, et al. Risk of osteoarthritis associated with long-term
weight-bearing sports. Arthritis Rheum 1996;39(6):988e95.
80. Sandmark H. Musculoskeletal dysfunction in physical
education teachers. Occup Environ Med 2000;57(10):673e7.
81. Kettunen JA, Kujala UM, Kaprio J, Koskenvuo M, Sarna S.
Lower-limb function among former elite male athletes. Am J
Sports Med 2001;29(1):2e8.
82. Jensen LK. Knee-straining work activities, self-reported knee
disorders and radiographically determined knee osteoarthritis.
Scand J Work Environ Health 2005;31:68e74.
83. Andrianakos AA, Kontelis LK, Karamitsos DG, Aslanidis SI,
Georgountzos AI, Kaziolas GO, et al. Prevalence of symptomatic
knee, hand, and hip osteoarthritis in Greece. The ESORDIG
study. J Rheumatol 2006;33(12):2507e13.
84. Tangtrakulwanich B, Chongsuvivatwong V, Geater AF. Associ-
ations between floor activities and knee osteoarthritis in Thai
Buddhist monks: the Songkhla study. J Med Assoc Thai
2006;89(11):1902e8.
85. Elleuch MH, Guermazi M, Mezghanni M, Ghroubi S, Fki H,
Mefteh S, et al. Knee osteoarthritis in 50 former top-level
soccer players: a comparative study. Ann Readapt Med Phys
2008;51(3):174e8.
86. Fernandez-Lopez JC, Laffon A, Blanco FJ, Carmona L,
Aretxabala I, Ballina J, et al. Prevalence, risk factors, and impact
of knee pain suggesting osteoarthritis in Spain. Clin Exp
Rheumatol 2008;26(2):324e32.
87. Muraki S, Akune T, Oka H, Mabuchi A, En-Yo Y, Yoshida M,
et al. Association of occupational activity with radiographic
knee osteoarthritis and lumbar spondylosis in elderly patients
of population-based cohorts: a large-scale population-based
study. Arthritis Rheum 2009;61(6):779e86.
88. Wang Y-Q, Wang X-H, Liu T-S, Fu Y, Lin J, Liu D- Y. Main
pathogenic factors in knee osteoarthritis patients, outpatients
and community staff from Xi’an, Tianjin and Guangzhou:
multicenter cluster sampling epidemiological survey. J Clin
Rehabil Tissue Eng Res 2009;13(41):8155e9 [Chinese].
89. Kim I, Kim HA, Seo YI, Song YW, Jeong JY, Kim DH. The prev-
alence of knee osteoarthritis in elderly community residents in
Korea. J Korean Med Sci 2010;25(2):293e8.
90. Bernard TE, Wilder FV, Aluoch M, Leaverton PE. Job-related
osteoarthritis of the knee, foot, hand, and cervical spine.
J Occup Environ Med 2010;52(1):33e8.