Professional Documents
Culture Documents
1 s2.0 S105687271600074X Main PDF
1 s2.0 S105687271600074X Main PDF
1 s2.0 S105687271600074X Main PDF
University of California San Francisco, Internal Medicine Residency, 505 Parnassus Avenue, San Francisco, CA 94143-0119
5 East 98th Street, Box 1188, New York, NY 10029
Cleveland Clinic, 9500 Euclid Avenue, Cleveland, OH 44195
a r t i c l e
i n f o
Article history:
Received 5 November 2015
Received in revised form 15 February 2016
Accepted 18 February 2016
Available online 2 March 2016
Keywords:
Type 2 diabetes
Obesity
Osteoarthritis
Meta-analysis
Prediabetes
a b s t r a c t
Background: Observational studies have reported an association between type 2 diabetes and osteoarthritis
(OA) development and progression. However no systematic review of the literature exists assessing whether
this association is consistently true. We aimed to systematically review the association between type 2
diabetes and the presence, development, and progression of OA.
Methods: We searched MEDLINE, SCOPUS, EMBASE, the Web of Science, and Grey Literature (through August
2014) for prospective cohort, cross-sectional, and casecontrol studies with condence intervals (CI) that
reported an association between type 2 diabetes and impaired glucose tolerance (IGT) and the development
or presence of OA of any joint.
Results: Ten studies and fourteen ratios were included in the analysis. The pooled population size in our
meta-regression was 16,742 patients. Type 2 diabetes was signicantly associated with the development or
presence of OA (OR; 1 21, 95% CI: 1 021 41). In the subset of 7 studies that did control for weight or BMI
there was an increased odds of OA associated with type 2 diabetes was (OR: 1 25, 95% CI: 1 051 46)
from a smaller pool of patients (n = 7156).
Conclusions: Type 2 diabetes is associated with the development and presence of radiographic and
symptomatic OA even when controlling for body mass index and weight.
2016 Elsevier Inc. All rights reserved.
1. Introduction
In the United States, osteoarthritis (OA) is estimated to affect
26 9 million Americans with a mean cost of over $2000 per patient
per year (Gabriel, Crowson, Campion, & OFallon, 1997; Lawrence,
Felson, Helmick, et al., 2008). OA is historically considered a
non-inammatory arthritis whose development is primarily associated with age and weight, but a growing body of literature has shed
light on the role of inammation and type 2 diabetes mellitus in the
development of the disease (Berenbaum, 2011). The independent
prevalence of both conditions has risen in the United States as the
population ages and struggles with the obesity epidemic (Deshpande,
2008; Gregg, Beckles, Williamson, et al., 2000; National Diabetes Fact
Sheet, 2011; Risk Factors for Non-Insulin Dependent Diabetes).
Among patients with type 2 diabetes, the prevalence of OA is
signicantly higher than in those without the condition (Cheng et al.,
2012; Puenpatom & Victor, 2009). Recent studies have found that the
Conicts of interest: None.
Funding: None.
Corresponding author at: Cleveland Clinic Main Campus, Mail Code F20, 9500 Euclid
Avenue, Cleveland, OH 44195. Tel.: +1 216 445 2679.
E-mail address: kashyas@ccf.org (S.R. Kashyap).
http://dx.doi.org/10.1016/j.jdiacomp.2016.02.016
1056-8727/ 2016 Elsevier Inc. All rights reserved.
M.F. Williams et al. / Journal of Diabetes and Its Complications 30 (2016) 944950
2. Methods
We followed Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines during the conduct of this systematic
review.
2.1. Data Sources and Searches
MEDLINE, SCOPUS, Web of Science, and EMBASE were searched in
July 2013 using optimally sensitive search strategies for relevant studies
with key search terms osteoarthritis and diabetes (and their acronyms
such as OA and DM). Authors were updated regarding publications
meeting the original search parameters through August 2014. We
reviewed the references of included studies for any further potential
studies. The authors of studies that assessed the association between OA
and type 2 diabetes but did not include a hazard ratio (HR), relative risk
(RR), or odds ratio (OR) were contacted for further information by
electronic and traditional mail. Inquiries were made as to whether
analyses pertinent to the association of interest were performed, as well
as details of their study design and patient population.
2.2. Study Selection
Two reviewers (M.F.W. and D.A.L.) independently analyzed the
search results and came to a consensus regarding which studies
contained pertinent data. CaseControl, cross-sectional, and prospective
cohort studies that reported the presence, development, or progression
of OA of weight bearing and non-weight bearing joints were included in
the study. Non-English language studies' abstracts were excluded.
Studies were included if the outcome was the presence or development
of OA of any joint. This was dened radiographically (Kellgren-Lawrence
Scale or Atlas of Rheumatology), clinically (American College of
Rheumatology criteria for diagnosis of OA), symptomatically (WOMAC
score), by outcome (requiring arthroplasty) or by diagnosis codes.
Studies that included hyperglycemic states including impaired fasting
glucose and type 2 diabetes as an independent variable were included if
it was dened by patient history, laboratory value denitions (i.e. fasting
945
946
M.F. Williams et al. / Journal of Diabetes and Its Complications 30 (2016) 944950
were pooled from individual studies. Only the adjusted risk estimates
that were most similar across studies were extracted and included in
the analysis. The pooled risk estimate was reported using the random
effects model because of anticipated statistical heterogeneity. A
sensitivity analysis was conducted using the risk estimates that
controlled for weight or BMI to assess whether statistical signicance
persisted in the presence of this potential confounder. Heterogeneity
was analyzed using a chi-squared test of N-1 degrees of freedom, with
the I2 test, and with an alpha of 0 05 dening statistical signicance
(Higgins, Thompson, Deeks, & Altman, 2003). Analyses were performed
using Comprehensive Meta-analysis (Biostat, Englewood, NJ) and
STATA/SE 10 1 (Stata Corporation, College Station, TX) software.
3. Results
3.1. Search Results
Our systematic review of the literature through August 2014
identied 1079 records that t our broad search strategy and a review
of their references resulted in an additional 42 abstracts of interest for
a total of 1121 abstracts (Fig. 1). After removal of duplicates and
reviewing the abstracts manually, 820 unique records remained. The
majority of records (n = 739) were excluded because only a title, but
no abstract or publication, could be obtained in English, or the
abstracts did not contain information pertaining to the disease entities
Table 1
Studies Included in Meta-Analysis.
Study
Cohort studies
Bagge et al.
(1991)
Sample size of
Recruitment source
study population
analyzed in
meta-analysis
340
Diabetes denition
Joint
involvement
Outcome
Group
analyzed
Pertinent
ratio
Factors
adjusted
Kellgren
Lawrence &
Symptoms
Total (n =
340)
Male
OA development
via Kellgren
Lawrence &
Symptoms
Kellgren
Lawrence &
Symptoms
Knee
replacement
Total
1.22
(0.722.07)
1.08
(0.512.28)
1.44
(0.543.00)
(0.351.14)
Total
(1.374.27)
Age, sex,
BMI
Total
(1.113.84)
Age, sex,
BMI, prior
joint
replacement
Knee
Kellgren
Lawrence 2 &
Symptoms
Female
Physician diagnosis
or use of oral
antihypergly-cemics
or insulin
Diagnosis by doctor
or on medications
Kellgren
Lawrence &
Symptoms
Symptomatic OA
conrmed with
ACR criteria &
Radiographs
Kellgren
Lawrence &
Symptoms
Female
Haugen et al.
(2011 abstract)
522
Framingham OA Study
Offspring Cohort, USA
Elevated blood
glucose
Yoshimura et al.
(2012)
1384
Schett et al.
(2013)
927
Questionnaire
regarding previous
diagnosis
Strmer et al.
(2001)
171
Andrianakos et al.
(2006)
547
Dahaghin et al.
(2007)
3585
Community cohort.
Erasmus Medical Centre,
Rotterdam
6299
CaseControl studies
Nieves-Plaza et al. 202
(2013)
Diagnosis Code
of OA
Southern Califronia
American Indian Health
Clinics Database
Diabetes diagnosed
by diagnosis code
Unclear based
on diagnosis
codes
Endocrinology clinics,
University of Puerto Rico's
Medical Sciences Campus
Plasma glucose
concentration,
fasting plasma,
2-hour plasma
1.02
(0.522.00)
Male
2.14
(0.924.96)
Generalized 0.97
OA
(0.50 1.86)
Age, sex,
BMI
Knee OR
1.26
(0.981.63)
Age, sex,
obesity
Total
1.20
(0.901.60)
1.90
(1.003.8)
Age
stratied,
sex, BMI
5562
years
62.168.7
years
N68.8
years
3544
years
4554
years
5564
years
N65 years
1.10
(0.701.8)
0.90
(0.601.4)
1.90
(1.063.40)
1.62
(1.092.42)
1.31
(0.811.99)
1.25
(0.801.94)
Total
2.18
(1.124.24)
Age
stratied, sex
Age, sex,
BMI
M.F. Williams et al. / Journal of Diabetes and Its Complications 30 (2016) 944950
947
Sarzi-Puttini, Scarpa, et al., 2005). The risk ratios from the Andrianakos
and Sturmer studies were updated following communication with those
authors. Finally, there were two publications from the ROAD study: the
one study that only included prevalence data was excluded, while the
publication with incidence data was included (Yoshimura et al., 2012).
Regarding those studies that assessed the association but did not
include a pertinent ratio, 2 found a statistically signicant positive
association between OA and type 2 diabetes (Puenpatom & Victor,
2009; Silveri, Brecciaroli, Argentati, & Cervini, 1994), 1 found a
statistically signicant negative association (Horn et al., 1992), 1
failed to show an association (Magar et al., 1989), while 1 varied
between a positive or no association depending on the joint analyzed
(Waine, Nevinny, Rosenthal, & Joffe, 1961). An additional 4 studies did
not analyze the association beyond prevalence in a population of
diabetics. The nal 2 studies discussed pertinent analyses, with one
failing to show a signicant statistical relationship while the other
found no signicant difference; however, the publications did not
discuss further details, and the authors could not be reached (Carman,
Sowers, Hawthorne, & Weissfeld, 1994; Frey, Barrett-Connor, Sledge,
Schneider, & Weisman, 1996). Notable amongst these 11 rejected
studies are two publications that were foundational in presenting the
question of an association between type 2 diabetes and OA
(Silberberg, Frank, Jarrett, & Silberberg, 1959; Waine et al., 1961).
Seventeen studies were included in the qualitative analysis but 5
were excluded from the nal quantitative synthesis. Two studies were
excluded as their primary outcome related to mortality (Gabriel,
Crowson, & OFallon, 1999; Nesch et al., 2011). A study from the
AMICA cohort had proxy outcomes of pain on the visual analog scale and
use of non-steroidal anti-inammatory drugs, while the Horn study
employed osteophytes as their outcome. Given that these outcomes did
not meet our inclusion criteria they were excluded (Cimmino,
Table 2
Assessment of Study Quality Using the Moose Criteria.
Cohort studies
Study
****
****
**
**
***
**
9
8
****
**
***
****
**
***
Cross-sectional studies
Study
Anderson and
Felson (1988)
Strmer et al.
(2001)
Andrianakos et al.
(2006)
Dahaghin et al.
(2007)
Reid et al. (2010)
***
**
**
**
**
***
**
**
**
CaseControl study
Study
Nieves-Plaza et al. **
(2013)
***
**
* indicates a number indicating study quality according to Moose Criteria. The greater
number of * indicates higher quality.
3.5. Meta-analysis
Using the random effects model, we found a statistically signicant
increased odds ratio of OA associated with the presence or
development of type 2 diabetes of 1 21 (95% CI 1 021 41)
from a pooled sample of 16,742 patients (Fig. 2). The heterogeneity
detected in these studies generated an I 2 value of 26% indicating mild
to moderate heterogeneity between the results of the studies. The
sensitivity analysis performed to assess whether the odds of the
association between type 2 diabetes and OA remained the same when
controlling for weight and BMI showed a consistent result to the
overall analysis.
948
M.F. Williams et al. / Journal of Diabetes and Its Complications 30 (2016) 944950
Study Name
Standard
Error
1.080
1.440
0.640
2.420
2.060
1.020
0.970
1.260
1.200
1.900
1.620
1.310
1.250
2.180
1.215
0.273
0.628
0.202
0.740
0.696
0.378
0.347
0.163
0.179
0.597
0.339
0.301
0.291
0.967
0.099
-2.00
-1.00
0.00
1.00
2.00
Favors
Association
M.F. Williams et al. / Journal of Diabetes and Its Complications 30 (2016) 944950
1.250
949
0.101
Fig. 3. Association between type 2 diabetes & OA in studies controlling for weight and bmi.
Acknowledgements
We thank Marian Simonson of the Cleveland Clinic Alumni Library
for her assistance in approaching the systematic review of the
literature.
References
Anderson, J. J., & Felson, D. T. (1988). Factors associated with osteoarthritis of the knee
in the rst national Health and Nutrition Examination Survey (HANES I). Evidence
for an association with overweight, race, and physical demands of work. American
Journal of Epidemiology, 128(1), 179189.
Andrianakos, A. A., Kontelis, L. K., Karamitsos, D. G., et al. (2006). Prevalence of
symptomatic knee, hand, and hip osteoarthritis in Greece. The ESORDIG study.
Journal of Rheumatology, 33(12), 25072513.
Bagge, E., Bjelle, A., Edn, S., & Svanborg, A. (1991). Factors associated with radiographic
osteoarthritis: Results from the population study 70-year-old people in Gteborg.
Journal of Rheumatology, 18(8), 12181222.
Berenbaum, F. (2011). Diabetes-induced osteoarthritis: From a new paradigm to a new
phenotype. Annals of the Rheumatic Diseases, 70(8), 13541356. http://dx.doi.org/
10.1136/ard.2010.146399.
Carman, W. J., Sowers, M., Hawthorne, V. M., & Weissfeld, L. A. (1994). Obesity as a risk
factor for osteoarthritis of the hand and wrist: A prospective study. American
Journal of Epidemiology, 139(2), 119129.
Cheng, Y. J., Imperatore, G., Caspersen, C. J., Gregg, E. W., Albright, A. L., & Helmick, C. G.
(2012). Prevalence of diagnosed arthritis and arthritis-attributable activity
limitation among adults with and without diagnosed diabetes: United States,
2008-2010. Diabetes Care, 35(8), 16861691. http://dx.doi.org/10.2337/dc12-0046.
Cimmino, M. A., Sarzi-Puttini, P., Scarpa, R., et al. (2005). Clinical presentation of
osteoarthritis in general practice: Determinants of pain in Italian patients in the
AMICA study. Seminars in Arthritis and Rheumatism, 35(1 Suppl. 1), 1723. http://
dx.doi.org/10.1016/j.semarthrit.2005.01.015.
Cooper, C., McAlindon, T., Snow, S., et al. (1994). Mechanical and constitutional risk
factors for symptomatic knee osteoarthritis: Differences between medial tibiofemoral and patellofemoral disease. Journal of Rheumatology, 21(2), 307313.
Dahaghin, S., Bierma-Zeinstra, S. M. A., Koes, B. W., Hazes, J. M. W., & Pols, H. A. P.
(2007). Do metabolic factors add to the effect of overweight on hand
osteoarthritis? The Rotterdam Study. Annals of the Rheumatic Diseases, 66(7),
916920. http://dx.doi.org/10.1136/ard.2005.045724.
DeGroot, J., Verzijl, N., Jacobs, K. M., et al. (2001). Accumulation of advanced glycation
endproducts reduces chondrocyte-mediated extracellular matrix turnover in
human articular cartilage. Osteoarthritis and Cartilage, 9(8), 720726. http://dx.
doi.org/10.1053/joca.2001.0469.
DeGroot, J., Verzijl, N., Wenting-van Wijk, M. J. G., et al. (2004). Accumulation of
advanced glycation end products as a molecular mechanism for aging as a risk
factor in osteoarthritis. Arthritis and Rheumatism, 50(4), 12071215. http://dx.doi.
org/10.1002/art.20170.
950
M.F. Williams et al. / Journal of Diabetes and Its Complications 30 (2016) 944950
Reid, J. L., Morton, D. J., Wingard, D. L., Garrett, M. D., von Muhlen, D., & Slymen, D.
(2010). Obesity and other cardiovascular disease risk factors and their association
with osteoarthritis in Southern California American Indians, 2002-2006. Ethnicity
and Disease, 20(4), 416422.
Risk Factors for Non-Insulin Dependent Diabetes (s). http://diabetes.niddk.nih.gov/dm/
pubs/america/pdf/chapter9.pdf (Accessed September 28, 2012)
Saudek, D. M., & Kay, J. (2003). Advanced glycation endproducts and osteoarthritis.
Current Rheumatology Reports, 5(1), 3340.
Schett, G., Kleyer, A., Perricone, C., et al. (2013). Diabetes is an independent predictor for
severe osteoarthritis: Results from a longitudinal cohort study. Diabetes Care, 36(2),
403409. http://dx.doi.org/10.2337/dc12-0924.
Silberberg, M., Frank, E. L., Jarrett, S. R., & Silberberg, R. (1959). Aging and osteoarthritis of
the human sternoclavicular joint. American Journal of Pathology, 35(4), 851865.
Silveri, F., Brecciaroli, D., Argentati, F., & Cervini, C. (1994). Serum levels of insulin in overweight
patients with osteoarthritis of the knee. Journal of Rheumatology, 21(10), 18991902.
Steenvoorden, M. M. C., Huizinga, T. W. J., Verzijl, N., et al. (2006). Activation of receptor
for advanced glycation end products in osteoarthritis leads to increased stimulation