SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Objective

To summarize the overall relative risk of knee osteoarthritis (OA) associated with body mass index, and to estimate the potential risk reduction due to the control of this risk factor.

Methods

Six electronic databases were searched up to July 2010. Relative risk was estimated using odds ratio (OR). A random-effects model was used to pool the results. Risk reduction was estimated using population-attributable risk percentage (PAR%), i.e., the proportion of knee OA that would have been avoided if obesity had not been present in the population. The percentage of obesity in different populations was obtained from the International Obesity Task Force.

Results

Forty-seven studies (446,219 subjects) were included in the meta-analysis, of which there were 14 cohort, 19 cross-sectional, and 14 case–control studies. The overall pooled ORs for overweight and obese individuals were 2.02 (95% confidence interval [95% CI] 1.84–2.22) and 3.91 (95% CI 3.32–4.56), respectively. Risk reduction in terms of PAR% for knee OA varied from 8% in China to 50% in the US, depending on the prevalence of overweight and obesity. The reduction was greater in severe symptomatic OA than in asymptomatic radiographic OA.

Conclusion

Obesity is a risk factor for many conditions, including knee OA. The benefit of modifying this risk factor may cause significant risk reduction of knee OA in the general population, especially in Western countries where obesity is prevalent.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Osteoarthritis (OA) is the most common type of arthropathy and a leading cause of pain, restricted mobility, and functional decline, particularly in older people. It is estimated that approximately 25% of people ages >55 years have persistent knee pain, 10% of whom report having painful disabling knee OA (1). OA contributes substantially to direct and indirect health costs. Currently, there is no cure and the benefits of pharmacologic treatments commonly used for its management are often outweighed by their side effects (2).

Development of knee OA is associated with heredity, age, sex, obesity, previous knee injury, occupational factors (e.g., kneeling and squatting), physical activity, and knee malalignment (3). Obesity is one of the greatest health risks facing the world today, and the prevalence has increased globally in recent decades. Moreover, populations in developed countries are living longer. As obesity is growing progressively in older age (4), the prevalence of obesity-related symptomatic knee OA is anticipated to increase and is threatening to become a major public health problem worldwide. OA is also a major cause of total joint replacement. More than 450,000 primary total knee arthroplasties were performed in the US in 2004, and the number is projected to increase to 3.48 million by 2030 (5). Consequently, the economic impact of OA on society is anticipated to increase in parallel.

Obesity is one of the most important modifiable risk factors for the development (6–12) and progression (6, 12–15) of knee OA. It is suggested that excess weight increases mechanical loading of joints, leading to increased cartilage degradation and subsequent failure of the entire joint (16). Another proposed mechanism is mediation through metabolic and hormonal factors that disadvantage the health of joint tissues (16). In the Framingham Knee OA study, weight loss of approximately 5.1 kg over a 10-year period was found to reduce the risk of incident knee OA by 54% (odds ratio [OR] 0.46, 95% confidence interval [95% CI] 0.24–0.86) in women (17). Recently, a meta-analysis of randomized controlled trials (RCTs) has shown weight loss to reduce physical disability in obese patients diagnosed with knee OA (18). However, despite convincing evidence from observational studies and RCTs of the causal effect of obesity in knee OA, the benefit of preventing obesity at the population level has yet to be assessed (19). This is of considerable importance in support of public health initiatives aimed at reducing the burden of disease in the population.

We therefore carried out this study to 1) estimate the overall relative risk (RR) of knee OA associated with obesity using a meta-analysis of observational studies, and 2) calculate the risk reduction of knee OA in different populations according to the overall RR estimate and the prevalence of obesity.

Significance & Innovations

  • Body mass index (BMI) is one of the modifiable risk factors for knee osteoarthritis (OA). However, despite convincing evidence on the causal effect of the elevated BMI in knee OA, the potential risk reduction of knee OA due to the control of overweight/obesity at the population level has not been quantified. This is of considerable importance in support of public health initiatives aimed at reducing the burden of the disease in the general population.

  • We found risk reduction in knee OA to vary from 8% in China to 60% in the US, depending on the prevalence of obesity. The reduction was greatest in severe symptomatic OA awaiting total joint replacement, followed by symptomatic knee OA and asymptomatic radiographic OA.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Literature search: data sources and search strategy.

A systematic search of literature published up to July 2010 in 6 English-indexed electronic databases was conducted. Ovid Medline (1950), EMBase (1980), and AMED (1985) databases were searched using a comprehensive search strategy (see Supplementary Appendix A, available in the online version of this article at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658), and a keyword search using the terms “knee osteoarthritis or knee pain” and “obesity or body mass index or risk factor” and “cohort or case-control or cross-sectional” was applied to PubMed, ISI Web of Science, and CINAHL. Additional studies were obtained from reference lists of relevant articles. No language restrictions were applied.

Study selection.

All relevant abstracts obtained from each database were exported to Endnote, version X4.0.2 (Thomson Reuters), and duplicate articles were excluded. Only observational studies in humans were included. Diseases other than OA were excluded.

Eligible articles were retrieved and reviewed. For inclusion, studies had to fulfill the following criteria: 1) cohort, cross-sectional, or case–control study evaluating the association between overweight or obesity and incidence/development, radiographic progression of knee OA, or knee pain; 2) use of body mass index (BMI) in kg/m2 as a measure of overweight or obesity; and 3) knee OA as the primary outcome of interest, which was defined using radiographs and clinical or physician-diagnosed OA. Where knee pain was the primary outcome of interest, the mean age of the study participants had to be ≥45 years to minimize the chance of knee pain due to other musculoskeletal conditions.

Data extraction.

Study design–specific data extraction forms were used to record extracted information on study characteristics (source of population, sample selection, total number of study participants, mean age or age range of participants, sex ratio, BMI mean/range, and country of study), whether BMI was a primary exposure variable, definition of OA, reported or calculated RRs (OR, RR, and hazard ratio [HR]) and their corresponding 95% CIs, confounding factors, and statistical methods performed. For studies with more than one article from the same study population, we selected the one providing the best evidence based on study design, whether BMI was the primary exposure or sample size, and/or date of publication, so that only one article per study was included. For example, we included results from a cohort article rather than case–control or cross-sectional analysis. If the study design was the same, articles reporting risk estimates from BMI as a primary exposure were included.

Quality assessment.

All of the articles were fully reviewed by one investigator (SGM) and validated by a second investigator (MH) for key study characteristics such as study design, setting, and outcome measures. In addition, a 10% random sample was extracted independently by MH for other characteristics to determine the disagreement. Double extractions were applied when the percent disagreement was more than 2% for each article. Consensus was achieved by discussion and the third investigator (WZ) was involved, if necessary. We followed criteria recommended by the Meta-Analysis of Observational Studies in Epidemiology (20) to assess the methodologic quality of the publications.

Statistical analysis.

BMI was classified using the World Health Organization (WHO) criteria (normal: <25 kg/m2, overweight: 25–29.9 kg/m2, and obese: ≥30 kg/m2). RR (OR, RR, or HR) of overweight and/or obese compared with normal weight was calculated and OR was used in this meta-analysis as a general term for all RR measures. For studies where WHO categories were not applied, we selected the category most similar to the WHO, or converted the RR obtained from a continuous scale to reflect overweight or obese BMI categories. For example, if the RR was measured as an OR per unit increase of the BMI, we multiplied the logarithm of the OR by 5 or 10 for overweight or obese, respectively. If the RR was measured as an OR per 5 kg/m2 unit increase of the BMI, we multiplied the logarithm of this OR by 1 or 2 to represent the overweight or obese categories, respectively.

Potential publication bias was examined by funnel plots and Egger's test (21). Forest plots were used to examine the distribution of the effect sizes. We assessed heterogeneity using Cochran's Q test (22) and the I2 statistic (23). Subgroup analyses were undertaken according to different study-level characteristics, such as study design, setting, definition of knee OA, sex, study region, etc. Dose- response data were extracted and analyzed separately, if available.

A random-effects model was used to pool results if studies were heterogeneous or if the reason of heterogeneity could not be reasonably identified; otherwise, a fixed-effects model was used (24). Analyses were undertaken using StatsDirect, version 2.7.8.

A meta–regression model was used to adjust for covariates, where study design as case–control (0,1), adjustment for potential confounders (0,1), overweight or obese (1, 2), hospital setting (0,1), and type of knee OA (radiographic OA, symptomatic OA, and OA requiring total knee replacement [TKR]) (1, 2, 3) were examined. SPSS, version 16, was used to carry out meta–regression analysis.

Population-attributable risk percentage (PAR%) was used to determine risk reduction. It is defined as the proportion of the disease that could be eliminated if the exposure was prevented in a population (25). The PAR% is calculated by (It − Iu)/It × 100%, where It = the incidence of a disease in the total population and Iu = the incidence of a disease in the unexposed population. Alternatively, it may be derived from the following formula: Pe (OR − 1)/[Pe (OR − 1) + 1] × 100% (25), where Pe = the current prevalence of obesity, if the RR (the OR in this case) and percentage of exposure (e.g., overweight/obesity) in the population are known. We used pooled ORs and the current prevalence of obesity to compute the PAR% for 3 knee OA phenotypes: radiographic, symptomatic, and OA requiring TKR. Population-specific obesity prevalence estimates were obtained from the International Obesity Task Force (26).

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Study characteristics.

The systematic literature search yielded 3,056 citations. Of these, 409 were potentially relevant. After subsequent abstract and full-article review, 81 articles met our inclusion criteria. A further 27 articles were excluded because they were either duplicate publications or lacked required data. One study (27) only reported OR without 95% CI and the attempt to contact the corresponding author was unsuccessful. Ultimately, 47 studies that examined the association between the development of knee OA and BMI were included in the meta-analysis (Figure 1). We also performed a separate meta-analysis on radiographic knee OA progression (n = 6) and knee pain (n = 7).

thumbnail image

Figure 1. Results of the systematic literature search (up to July 1, 2010). * Knee osteoarthritis (OA) development = 41, knee OA incidence + radiographic progression = 4, radiographic knee OA progression = 2, knee OA development + knee pain = 2, knee pain only = 5. BMI = body mass index.

Download figure to PowerPoint

BMI and development of knee OA.

Of 47 studies (446,219 subjects) for the development of knee OA, there were 14 (6–12, 15, 28–33) cohort (404,070 subjects), 19 (34–52) cross-sectional (31,041 subjects), and 14 (53–66) case–control (11,108 subjects) studies (Table 1). Cohort studies had younger age, fewer women, and most (93%) examined BMI as a primary exposure. Details for each study can be found in Supplementary Appendix B (available in the online version of this article at http://online library.wiley.com/journal/10.1002/(ISSN)2151-4658).

Table 1. Characteristics of observational studies evaluating BMI and the development of knee OA*
 CohortCross-sectionalCase–controlAll studies
  • *

    BMI = body mass index; OA = osteoarthritis; TKR = total knee replacement.

  • Estimate excludes Sahlstrom et al (62), whose proportion of women was not stated.

  • Includes symptomatic radiographic OA, clinical OA (American College of Rheumatology), and self-reported OA.

  • §

    Includes incidence TKR due to OA and cases recruited from TKR waiting lists.

No. of studies14191447
No. of subjects404,07031,04111,108446,219
Age, range years20–8435–9047–9620–96
Women, no. (%)49,679 (12)17,894 (58)5,789 (56)73,362 (17)
No. of hospital-based studies1135
OA definition, no. (%)    
 Radiographic OA7 (50)15 (79)22 (47)
 Symptomatic OA4 (29)4 (21)7 (50)15 (32)
 OA requiring TKR§3 (21)7 (50)10 (21)
BMI as a primary exposure, no. (%)13 (93)16 (84)8 (57)37 (79)
Study region, no. (%)    
 US4 (29)6 (32)2 (14)12 (26)
 Europe8 (57)4 (21)10 (71)22 (47)
 Others2 (14)9 (47)2 (14)13 (28)

An asymmetric funnel plot showed significant publication bias; smaller studies with larger ORs were more likely to be published (Egger's test bias = 2.96, P = 0.0001) (Figure 2). There was considerable heterogeneity, with I2 of 92.9% across all 47 studies, with a pooled OR of 2.78 (95% CI 2.45–3.15) for overweight/obese (Figure 3).

thumbnail image

Figure 2. A funnel plot showing the association between overweight/obesity and the development of knee osteoarthritis in all 47 studies. I2 (inconsistency) = 92.9%; Egger's test bias = 2.96 (95% confidence interval 1.57–4.35); P = 0.0001.

Download figure to PowerPoint

thumbnail image

Figure 3. A forest plot showing the association between body mass index and knee osteoarthritis in all 47 studies.

Download figure to PowerPoint

Pooled ORs varied according to study design. Case–control studies had a greater OR (3.90, 95% CI 3.09–4.91) than cohort (2.51, 95% CI 2.06–3.07) and cross-sectional studies (2.41, 95% CI 2.00–2.91) (Table 2).

Table 2. Subgroup analysis for the association between BMI and knee OA*
 No. of studiesNo. of subjectsI2, % (Cochran's Q P)Random-effects pooled OR (95% CI)Publication bias (Egger's test bias P)
  • *

    BMI = body mass index; OA = osteoarthritis; OR = odds ratio; 95% CI = 95% confidence interval; TKR = total knee replacement.

  • Estimates compare specified BMI category with normal weight BMI category.

  • Assessment of BMI in 2 studies (56, 62) not known.

  • §

    Overweight/obese compared with normal weight BMI category.

By BMI categories     
 Underweight5328,9240.0 (0.637)0.78 (0.62–0.99)0.711
 Overweight35437,14078.3 (< 0.001)2.02 (1.84–2.22)0.020
 Obese37439,69386.3 (< 0.001)3.91 (3.32–4.56)0.037
 Overweight/obese47446,21992.9 (< 0.001)2.78 (2.45–3.15)< 0.001
By study design     
 Cohort     
  Overweight12402,53685.1 (< 0.001)1.94 (1.63–2.31)0.358
  Obese11401,92888.6 (< 0.001)3.17 (2.40–4.19)0.700
  Overweight/obese14404,07093.1 (< 0.001)2.51 (2.06–3.07)0.109
 Cross-sectional     
  Overweight1526,31031.4 (0.117)1.81 (1.66–1.97)0.390
  Obese1729,28179.3 (< 0.001)3.64 (2.96–4.47)0.106
  Overweight/obese1931,04192.6 (< 0.001)2.41 (2.00–2.91)0.018
 Case–control     
  Overweight88,29477.1 (< 0.001)2.66 (2.00–3.55)0.572
  Obese98,48474.2 (< 0.001)6.01 (4.12–8.77)0.506
  Overweight/obese1411,10872.2 (< 0.001)3.90 (3.09–4.91)0.701
By sex     
 Male     
  Overweight13338,87158.9 (0.004)2.21 (1.82–2.70)0.463
  Obese12338,16767.3 (< 0.001)4.00 (2.98–5.38)0.547
  Overweight/obese16341,18877.3 (< 0.001)2.79 (2.26–3.44)0.420
 Female     
  Overweight1730,84762.2 (< 0.001)2.09 (1.80–2.42)0.071
  Obese1731,01881.0 (< 0.001)4.26 (3.28–5.55)0.083
  Overweight/obese2133,02986.6 (< 0.001)3.14 (2.53–3.89)0.011
By study setting     
 Community43442,34092.8 (< 0.001)2.70 (2.38–3.08)< 0.001
 Hospital43,87989.4 (< 0.001)3.53 (2.19–5.69)0.566
By assessment of BMI     
 Measured BMI36427,47192.2 (< 0.001)2.69 (2.34–3.08)0.004
 Self-reported BMI916,25995.0 (< 0.001)3.11 (1.99–4.86)0.026
By OA symptoms and phenotypes     
 Radiographic knee OA     
  Overweight1826,40630.0 (0.112)1.85 (1.69–2.03)0.219
  Obese1826,55780.7 (< 0.001)3.36 (2.74–4.13)0.218
  Overweight/obese2229,29992.1 (< 0.001)2.46 (2.06–2.94)0.007
 Symptomatic radiographic OA     
  Overweight1018,57683.1 (< 0.001)1.91 (1.51–2.41)0.052
  Obese1220,95885.4 (< 0.001)3.98 (2.77–5.71)0.010
  Overweight/obese1523,38590.7 (< 0.001)2.66 (2.03–3.48)0.002
 Knee OA requiring TKR     
  Overweight7392,15885.6 (< 0.001)2.53 (2.02–3.19)0.034
  Obese7392,15888.1 (< 0.001)5.49 (3.87–7.80)0.137
  Overweight/obese10393,53593.9 (< 0.001)3.99 (2.83–5.64)0.010
By study location§     
 US1218,59384.1 (< 0.001)2.81 (2.25–3.50)0.050
 Europe22376,43694.4 (< 0.001)3.09 (2.41–3.96)< 0.001
 Others1351,19093.4 (< 0.001)2.37 (1.90–2.97)0.292

The results were still highly heterogeneous even after stratification by study design. Egger's test revealed a statistically significant publication bias in cross-sectional studies (P = 0.018) but not in cohort or case–control studies. The ORs did not differ significantly between sexes, study setting, assessment of BMI, or study location (Table 2).

A dose-response relationship was seen between BMI and knee OA, regardless of the definition of the disease. The pooled OR for overweight was 2.02 (95% CI 1.84–2.22), and for obese was 3.91 (95% CI 3.32–4.56) as compared with normal weight. We also examined the association between underweight and knee OA in comparison with normal weight. A marginal negative association (OR 0.78, 95% CI 0.62–0.99) was found (Table 2).

Meta-regression.

Of the 4 covariates included in this analysis, only case–control studies (OR 1.53, 95% CI 1.07–2.19) and obesity (OR 1.84, 95% CI 1.40–2.40) were statistically significant.

PAR%.

Using the pooled OR for radiographic, symptomatic, and knee OA requiring TKR (Table 2) and the country-specific overweight/obesity prevalence, the PAR% was calculated (Table 3). For example, the risk reduction of symptomatic knee OA due to the prevention of overweight varied from 13% in China to 24% in the US, while that for obesity ranged from 8% to 50%, respectively. This variation did not only depend on the prevalence of overweight or obesity, but also on the severity of knee OA. The PAR% was smaller in asymptomatic radiographic OA (data not reported), but greater in severe knee OA requiring TKR (Table 3).

Table 3. Risk reduction (PAR%) of knee OA for overweight and obesity in selected countries*
CountryOverweight (BMI 25–29.9 kg/m2)Obesity (BMI ≥30 kg/m2)
PrevalenceSymptomatic knee OA, PAR% (95% CI)Knee OA requiring TKR, PAR% (95% CI)§PrevalenceSymptomatic knee OA, PAR% (95% CI)Knee OA requiring TKR, PAR% (95% CI)§
  • *

    PAR% = population-attributable risk percentage; OA = osteoarthritis; BMI = body mass index; 95% CI = 95% confidence interval; TKR = total knee replacement.

  • Prevalence estimates were obtained from the International Obesity Task Force (26).

  • Pooled odds ratios (ORs) 1.91 (95% CI 1.51–2.41) for overweight and 3.98 (95% CI 2.77–5.71) for obesity associated with symptomatic knee OA (Table 2) were used.

  • §

    Pooled ORs 2.53 (95% CI 2.02–3.19) for overweight and 5.49 (95% CI 3.87–7.80) for obesity associated with knee OA requiring TKR (Table 2) were used.

  • Average obesity prevalence from the UK.

US34.223.7 (14.9–32.5)34.4 (25.9–42.8)33.850.2 (37.4–61.4)60.3 (49.2–69.7)
Australia36.524.9 (15.7–34.0)35.8 (27.1–44.4)24.842.5 (30.5–53.9)52.7 (41.6–62.8)
UK36.925.1 (15.8–34.2)36.1 (27.3–44.7)24.542.2 (30.2–53.6)52.4 (41.3–62.5)
Germany37.525.4 (16.1–34.6)36.5 (27.7–45.1)20.838.3 (26.9–49.5)48.3 (37.4–58.6)
Norway46.629.8 (19.2–39.7)41.6 (32.2–50.5)18.335.2 (24.4–46.2)45.0 (34.4–55.4)
Morocco27.720.1 (12.4–38.1)29.7 (22.0–37.7)15.030.8 (20.9–41.3)40.2 (30.0–50.4)
Spain38.626.0 (16.4–35.2)37.1 (28.2–45.8)14.630.3 (20.5–40.7)39.6 (29.5–49.8)
Finland35.824.5 (15.4–33.5)35.4 (26.7–43.9)14.229.7 (20.1–40.1)38.9 (29.0–49.1)
Sweden35.124.2 (15.2–33.1)34.9 (26.3–43.4)12.927.8 (18.6–37.8)36.7 (27.0–46.7)
The Netherlands36.024.7 (15.5–33.7)35.5 (26.9–44.1)10.323.4 (15.4–32.6)31.5 (22.7–41.1)
Thailand21.516.3 (9.9–23.2)24.7 (18.0–32.0)6.917.1 (10.9–24.5)23.7 (16.5–31.9)
China16.112.7 (7.6–18.5)19.7 (14.1–26.0)2.98.0 (4.9–12.0)11.5 (7.7–16.5)

Others.

BMI and radiographic progression of knee OA.

Six studies (5,778 subjects) (6, 11–15) investigated the association between BMI and radiographic progression of knee OA (see Supplementary Appendix C, available in the online version of this article at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658). Of these, 3 studies (13–15) found a positive significant association with overweight and 4 studies (12–15) with obese categories. The pooled ORs were 1.33 (95% CI 1.01–1.75) for overweight and 1.93 (95% CI 1.23–3.04) for obese.

BMI and knee pain.

Seven studies (18,450 subjects) (45, 46, 49, 67–70) examined the association between BMI and persistent knee pain (see Supplementary Appendix D, available in the online version of this article at http://onlinelibrary.wiley.com/journal/10.1002/(ISSN)2151-4658). Of these, 5 studies (46, 49, 67–69) showed a statistically significant association between overweight and knee pain, whereas all of the 7 studies (45, 46, 49, 67–70) found a significant association in the obese category. The pooled ORs were 1.69 (95% CI 1.36–2.11) and 2.58 (95% CI 1.74–3.82) for the overweight and obese BMI categories, respectively.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

BMI is a major risk factor for the development of knee OA (3, 71), and it may also accelerate the progression of the disease (72). This meta-analysis of 47 observational studies with more than 446,000 people found an approximately 3-fold increased risk in the development of knee OA in overweight or obese individuals, and the association was dose dependent. A potential risk reduction may be estimated according to these RRs estimated from meta-analysis. For example, in the US, approximately 50% of symptomatic knee OA may be prevented if the current obesity prevalence in the US population could be controlled. The risk reduction in China, in contrast, is only 8% due to the low prevalence of obesity (Table 3). It is predicted, however, that the prevalence of obesity is increasing dramatically (73, 74). The impact of preventing obesity and the resultant risk reduction are substantial, especially when multiple outcomes of obesity are considered (19).

To our knowledge, this is the first calculation of risk reduction using the RR summarized from a meta-analysis. The advantage of this method is that it first establishes the association between exposure and the condition. Second, it produces a powerful and globally generalizable RR estimate for the estimation of risk reduction. Given a country- or population-specific prevalence of exposure (e.g., obesity), users are able to estimate the risk reduction for their own populations. The utility of the RR estimates provided in Table 2 is significant. We have therefore listed all possible ORs in subgroups to accommodate the application. The country-specific PAR% in Table 3 is based on the current prevalence of overweight and obesity in adults. It is important to note that we have used prevalence data of adult obesity that includes young adults. However, since BMI increases with age while OA is associated with older age, PAR% for obesity may be underestimated; therefore, these estimates are for reference only. Moreover, risk reduction due to the prevention of both overweight and obesity may be additive. Therefore, the incidence of knee OA would be greatly reduced if overweight and obesity were eliminated.

This study has several limitations. First, we combined RR estimates from different study designs for the subsequent calculation of PAR%. Yet, the OR from case–control studies is less conservative and can overestimate the RR in common diseases such as knee OA. Moreover, the OR may change with the number of confounding factors being adjusted for. As this was a study-level analysis, we were unable to provide an OR that was adjusted with the same set of confounding factors, and unadjusted ORs were used where adjustment was unavailable. Therefore, the RR may be overestimated. Second, there was a high degree of variation among studies included in the meta-analysis. We carried out a number of subgroup analyses with regard to potential sources of heterogeneity such as study design, setting (hospital versus community), sex, adjustment for confounding factors, disease definition, assessment of BMI (self-reported or measured), and BMI categories. Even so, we were unable to explain some of the heterogeneity found here, but findings from the meta–regression analysis revealed case–control studies and obesity as the main sources. Third, it is conceivable that persons may gain weight after developing knee OA as a consequence of knee pain and physical inactivity. Therefore, cross-sectional BMI assessed in symptomatic knee OA cases may not reflect the body weight prior to the onset of knee OA. This may be a source of bias and may also explain the high pooled risk estimate in case–control studies. Finally, we assessed BMI as a surrogate measure of adiposity. However, BMI neither distinguishes heaviness due to body fat or heaviness due to muscle bulk, nor does it indicate the distribution of body fat. It is speculated that fat distribution may act through metabolic or systematic rather than biomechanical mechanisms to contribute to knee OA. Some epidemiologic studies have demonstrated that certain adiposity measures such as waist circumference and waist to hip ratio or body composition components such as fat mass may be associated with knee OA (31, 34, 43, 75), but not others (36, 76). In spite of these reports, such adiposity measures were beyond the scope of this analysis.

Evidence suggests that excessive body weight may attenuate biomechanical load exerted on weight-bearing joints. It is also plausible that obesity may additionally cause dysfunctional metabolism and joint damage by stimulating adipokines that are known regulators of metabolic homeostasis (77, 78). This may partly explain the different associations between obesity and the incidence of OA in weight-bearing versus non–weight-bearing joints such as hand OA (29, 79). Weight loss has shown significant clinical effects in the treatment of OA in a number of RCTs for weight-bearing joint OA (2). However, to our knowledge there have been no trials to prove whether weight loss can prevent the onset of knee OA. Therefore, we have summarized evidence from the observational studies to examine the potential benefit of preventing knee OA attributable to overweight and obesity. It was not our intention to investigate the effectiveness of weight reduction interventions, but to provide an epidemiologic perspective on the potential risk reduction of knee OA in a general population if overweight/obesity was removed. However, these estimates may carry confounding bias, and further prevention trials on the effect of weight reduction are needed to confirm these results.

In summary, obesity is a common risk factor associated with many conditions, including knee OA. It is one of the few risk factors that may be modified, although effective modification is challenging. The resulting benefits from preventing obesity in the general population are huge, especially in the Western world, where obesity is highly prevalent. Approximately half of symptomatic knee OA, for example, would be prevented in the US if this risk factor was eliminated.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Zhang had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Muthuri, Doherty, Zhang.

Acquisition of data. Muthuri, Hui.

Analysis and interpretation of data. Muthuri, Doherty, Zhang.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

We would like to thank Joanna Ramowski for data collection, Helen Richardson for logistics support, and Maggie Wheeler for language translations.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information
  • 1
    Peat G, McCarney R, Croft P. Knee pain and osteoarthritis in older adults: a review of community burden and current use of primary health care. Ann Rheum Dis 2001; 60: 917.
  • 2
    Zhang W, Nuki G, Moskowitz RW, Abramson S, Altman RD, Arden NK, et al. OARSI recommendations for the management of hip and knee osteoarthritis: part III. Changes in evidence following systematic cumulative update of research published through January 2009. Osteoarthritis Cartilage 2010; 18: 47699.
  • 3
    Felson DT, Lawrence RC, Dieppe PA, Hirsch R, Helmick CG, Jordan JM, et al. Osteoarthritis: new insights. Part 1: the disease and its risk factors. Ann Intern Med 2000; 133: 63546.
  • 4
    Rossner S. Obesity in the elderly: a future matter of concern? Obes Rev 2001; 2: 1838.
  • 5
    Kurtz S, Ong K, Lau E, Mowat F, Halpern M. Projections of primary and revision hip and knee arthroplasty in the United States from 2005 to 2030. J Bone Joint Surg Am 2007; 89: 7805.
  • 6
    Cooper C, Snow S, McAlindon TE, Kellingray S, Stuart B, Coggon D, et al. Risk factors for the incidence and progression of radiographic knee osteoarthritis. Arthritis Rheum 2000; 43: 9951000.
  • 7
    Hart DJ, Doyle DV, Spector TD. Incidence and risk factors for radiographic knee osteoarthritis in middle-aged women: the Chingford study. Arthritis Rheum 1999; 42: 1724.
  • 8
    Hochberg MC, Lethbridge-Cejku M, Tobin JD. Bone mineral density and osteoarthritis: data from the Baltimore Longitudinal Study of Aging. Osteoarthritis Cartilage 2004; 12: S458.
  • 9
    Manninen P, Riihimaki H, Heliovaara M, Makela P. Overweight, gender and knee osteoarthritis. Int J Obes Relat Metab Disord 1996; 20: 5957.
  • 10
    McAlindon T, Zhang Y, Hannan M, Naimark A, Weissman B, Castelli W, et al. Are risk factors for patellofemoral and tibiofemoral knee osteoarthritis different? J Rheumatol 1996; 23: 3327.
  • 11
    Niu J, Zhang YQ, Torner J, Nevitt M, Lewis CE, Aliabadi P, et al. Is obesity a risk factor for progressive radiographic knee osteoarthritis? Arthritis Rheum 2009; 61: 32935.
  • 12
    Reijman M, Pols HA, Bergink AP, Hazes JM, Belo JN, Lievense AM, et al. Body mass index associated with onset and progression of osteoarthritis of the knee but not of the hip: the Rotterdam study. Ann Rheum Dis 2007; 66: 15862.
  • 13
    Lachance L, Sowers MF, Jamadar D, Hochberg M. The natural history of emergent osteoarthritis of the knee in women. Osteoarthritis Cartilage 2002; 10: 84954.
  • 14
    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: 9327.
  • 15
    Shiozaki H, Koga Y, Omori G, Tamaki M. Obesity and osteoarthritis of the knee in women: results from the Matsudai Knee Osteoarthritis Survey. Knee 1999; 6: 18992.
  • 16
    Felson DT. Does excess weight cause osteoarthritis and, if so, why? Ann Rheum Dis 1996; 55: 66870.
  • 17
    Felson DT, Zhang Y, Anthony JM, Naimark A, Anderson JJ. Weight loss reduces the risk for symptomatic knee osteoarthritis in women: the Framingham study. Ann Intern Med 1992; 116: 5359.
  • 18
    Christensen R, Bartels EM, Astrup A, Bliddal H. Effect of weight reduction in obese patients diagnosed with knee osteoarthritis: a systematic review and meta-analysis. Ann Rheum Dis 2007; 66: 4339.
  • 19
    Zhang W. Risk factors of knee osteoarthritis: excellent evidence but little has been done [editorial]. Osteoarthritis Cartilage 2010; 18: 12.
  • 20
    Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al, for the Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group. Meta-analysis of observational studies in epidemiology: a proposal for reporting. JAMA 2000; 283: 200812.
  • 21
    Egger M, Davey Smith G, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ 1997; 315: 62934.
  • 22
    Gavaghan DJ, Moore RA, McQuay HJ. An evaluation of homogeneity tests in meta-analyses in pain using simulations of individual patient data. Pain 2000; 85: 41524.
  • 23
    Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ 2003; 327: 55760.
  • 24
    Whitehead A, Whitehead J. A general parametric approach to the meta-analysis of randomized clinical trials. Stat Med 1991; 10: 166577.
  • 25
    Taylor JW. Simple estimation of population attributable risk from case-control studies. Am J Epidemiol 1977; 106: 260.
  • 26
    International Obesity Task Force. Global prevalence of adult obesity. URL: http://www.iotf.org/database/index.asp.
  • 27
    Li NH, Zhang YN, Zhang Y, Wang KZ, Li E, Zhu HM, et al. Comparison of the risk factors for knee osteoarthritis in middle-aged and elderly populations from six cities of six administration regions of China (Chinese). J Clin Rehabil Tissue Eng Res 2007; 11: 775860.
  • 28
    Gelber AC, Hochberg MC, Mead LA, Wang NY, Wigley FM, Klag MJ. Body mass index in young men and the risk of subsequent knee and hip osteoarthritis. Am J Med 1999; 107: 5428.
  • 29
    Grotle M, Hagen KB, Natvig B, Dahl FA, Kvien TK. Obesity and osteoarthritis in knee, hip and/or hand: an epidemiological study in the general population with 10 years follow-up. BMC Musculoskelet Disord 2008; 9: 132.
  • 30
    Jarvholm B, Lewold S, Malchau H, Vingard E. Age, bodyweight, smoking habits and the risk of severe osteoarthritis in the hip and knee in men. Eur J Epidemiol 2005; 20: 53742.
  • 31
    Lohmander LS, De Verdier MG, Rollof J, Nilsson PM, Engstrom G. Incidence of severe knee and hip osteoarthritis in relation to different measures of body mass: a population-based prospective cohort study. Ann Rheum Dis 2009; 68: 4906.
  • 32
    Toivanen AT, Heliovaara M, Impivaara O, Arokoski JP, Knekt P, Lauren H, et al. Obesity, physically demanding work and traumatic knee injury are major risk factors for knee osteoarthritis: a population-based study with a follow-up of 22 years. Rheumatology (Oxford) 2010; 49: 30814.
  • 33
    Wang Y, Simpson JA, Wluka AE, Teichtahl AJ, English DR, Giles GG, et al. Relationship between body adiposity measures and risk of primary knee and hip replacement for osteoarthritis: a prospective cohort study. Arthritis Res Ther 2009; 11: R31.
  • 34
    Abbate LM, Stevens J, Schwartz TA, Renner JB, Helmick CG, Jordan JM. Anthropometric measures, body composition, body fat distribution, and knee osteoarthritis in women. Obesity (Silver Spring) 2006; 14: 127481.
  • 35
    Al-Arfaj AS. Radiographic osteoarthritis and obesity. Saudi Med J 2002; 23: 93842.
  • 36
    Anderson JJ, Felson DT. Factors associated with osteoarthritis of the 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: 17989.
  • 37
    Aoda H, Nakamura K, Omori G, Koga Y, Akazawa K, Yamamoto M. Independent predictors of knee osteoarthritis in an elderly Japanese population: a multivariate analysis. Acta Med Biol (Niigata) 2006; 54: 3341.
  • 38
    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: 121822.
  • 39
    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: 338.
  • 40
    Cicuttini FM, Spector T, Baker J. Risk factors for osteoarthritis in the tibiofemoral and patellofemoral joints of the knee. J Rheumatol 1997; 24: 11647.
  • 41
    Du H, Chen SL, Bao CD, Wang XD, Lu Y, Gu YY, et al. Prevalence and risk factors of knee osteoarthritis in Huang-Pu District, Shanghai, China. Rheumatol Int 2005; 25: 58590.
  • 42
    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: 32432.
  • 43
    Janssen I, Mark AE. Separate and combined influence of body mass index and waist circumference on arthritis and knee osteoarthritis. Int J Obes 2006; 30: 12238.
  • 44
    Kim I, Kim HA, Seo YI, Song YW, Jeong JY, Kim DH. The prevalence of knee osteoarthritis in elderly community residents in Korea. J Korean Med Sci 2010; 25: 2938.
  • 45
    Lachance L, Sowers M, Jamadar D, Jannausch M, Hochberg M, Crutchfield M. The experience of pain and emergent osteoarthritis of the knee. Osteoarthritis Cartilage 2001; 9: 52732.
  • 46
    Muraki S, Oka H, Akune T, Mabuchi A, En-yo Y, Yoshida M, et al. Prevalence of radiographic knee osteoarthritis and its association with knee pain in the elderly of Japanese population-based cohorts: the ROAD study. Osteoarthritis Cartilage 2009; 17: 113743.
  • 47
    Sudo A, Miyamoto N, Horikawa K, Urawa M, Yamakawa T, Yamada T, et al. Prevalence and risk factors for knee osteoarthritis in elderly Japanese men and women. J Orthop Sci 2008; 13: 4138.
  • 48
    Tangtrakulwanich B, Geater AF, Chongsuvivatwong V. Prevalence, patterns, and risk factors of knee osteoarthritis in Thai monks. J Orthop Sci 2006; 11: 43945.
  • 49
    Tukker A, Visscher TL, Picavet HS. Overweight and health problems of the lower extremities: osteoarthritis, pain and disability. Public Health Nutr 2009; 12: 35968.
  • 50
    Von Muhlen D, Morton D, Von Muhlen CA, Barrett-Connor E. Postmenopausal estrogen and increased risk of clinical osteoarthritis at the hip, hand, and knee in older women. J Womens Health Gend Based Med 2002; 11: 5118.
  • 51
    Wang W, Wang KZ, Dang XQ, Bai CY, Li NH, Wang CS, et al. Relevant factors for knee osteoarthritis in middle aged and old population. Chin J Clin Rehabil 2006; 10: 158. In Chinese.
  • 52
    Zeng QY, Zang CH, Li XF, Dong HY, Zhang AL, Lin L. Associated risk factors of knee osteoarthritis: a population survey in Taiyuan, China. Chin Med J (Engl) 2006; 119: 15227.
  • 53
    Coggon D, Croft P, Kellingray S, Barrett D, McLaren M, Cooper C. Occupational physical activities and osteoarthritis of the knee. Arthritis Rheum 2000; 43: 14439.
  • 54
    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: 82330.
  • 55
    Holmberg S, Thelin A, Thelin N. Knee osteoarthritis and body mass index: a population-based case-control study. Scand J Rheumatol 2005; 34: 5964.
  • 56
    Klussmann A, Gebhardt H, 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.
  • 57
    Kohatsu ND, Schurman DJ. Risk factors for the development of osteoarthrosis of the knee. Clin Orthop Relat Res 1990; 261: 2426.
  • 58
    Limer KL, Tosh K, Bujac SR, McConnell R, Doherty S, Nyberg F, et al. Attempt to replicate published genetic associations in a large, well-defined osteoarthritis case-control population (the GOAL study). Osteoarthritis Cartilage 2009; 17: 7829.
  • 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: 2532.
  • 60
    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: 3236.
  • 61
    Oliveria SA, Felson DT, Cirillo PA, Reed JI, Walker AM. Body weight, body mass index, and incident symptomatic osteoarthritis of the hand, hip, and knee. Epidemiology 1999; 10: 1616.
  • 62
    Sahlstrom A, Johnell O, Redlund-Johnell I. The natural course of arthrosis of the knee. Clin Orthop Relat Res 1997; 340: 1527.
  • 63
    Sandmark H, Hogstedt C, Lewold S, Vingard E. Osteoarthrosis of the knee in men and women in association with overweight, smoking, and hormone therapy. Ann Rheum Dis 1999; 58: 1515.
  • 64
    Soeroso J, Dans LF, Amarillo ML, Santoso GH, Kalim H. Risk factors of symptomatic osteoarthritis of the knee at a hospital in Indonesia. APLAR J Rheumatol 2005; 8: 10613.
  • 65
    Sutton AJ, Muir KR, Mockett S, Fentem P. A case-control study to investigate the relation between low and moderate levels of physical activity and osteoarthritis of the knee using data collected as part of the Allied Dunbar National Fitness Survey. Ann Rheum Dis 2001; 60: 75664.
  • 66
    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: 291300.
  • 67
    Davis MA, Ettinger WH, Neuhaus JM, Barclay JD, Segal MR. Correlates of knee pain among US adults with and without radiographic knee osteoarthritis. J Rheumatol 1992; 19: 194349.
  • 68
    Miranda H, Viikari-Juntura E, Martikainen R, Riihimaki H. A prospective study on knee pain and its risk factors. Osteoarthritis Cartilage 2002; 10: 62330.
  • 69
    Sobti A, Cooper C, Inskip H, Searle S, Coggon D. Occupational physical activity and long-term risk of musculoskeletal symptoms: a national survey of post office pensioners. Am J Ind Med 1997; 32: 7683.
  • 70
    Jinks C, Jordan KP, Blagojevic M, Croft P. Predictors of onset and progression of knee pain in adults living in the community: a prospective study. Rheumatology (Oxford) 2008; 47: 36874.
  • 71
    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 2009; 18: 2433.
  • 72
    Belo JN, Berger MY, Reijman M, Koes BW, Bierma-Zeinstra SM. Prognostic factors of progression of osteoarthritis of the knee: a systematic review of observational studies. Arthritis Rheum 2007; 57: 1326.
  • 73
    Flegal KM, Carroll MD, Ogden CL, Curtin LR. Prevalence and trends in obesity among US adults, 1999-2008. JAMA 2010; 303: 23541.
  • 74
    Zaninotto P, Head J, Stamatakis E, Wardle H, Mindell J. Trends in obesity among adults in England from 1993 to 2004 by age and social class and projections of prevalence to 2012. J Epidemiol Community Health 2009; 63: 1406.
  • 75
    Hochberg MC, Lethbridge-Cejku M, Scott WW Jr, Reichle R, Plato CC, Tobin JD. The association of body weight, body fatness and body fat distribution with osteoarthritis of the knee: data from the Baltimore Longitudinal Study of Aging. J Rheumatol 1995; 22: 48893.
  • 76
    Davis MA, Neuhaus JM, Ettinger WH, Mueller WH. Body fat distribution and osteoarthritis. Am J Epidemiol 1990; 132: 7017.
  • 77
    Dumond H, Presle N, Terlain B, Mainard D, Loeuille D, Netter P, et al. Evidence for a key role of leptin in osteoarthritis. Arthritis Rheum 2003; 48: 311829.
  • 78
    Loeser RF. Systemic and local regulation of articular cartilage metabolism: where does leptin fit in the puzzle? [editorial]. Arthritis Rheum 2003; 48: 300912.
  • 79
    Carman WJ, Sowers M, Hawthorne VM, Weissfeld LA. Obesity as a risk factor for osteoarthritis of the hand and wrist: a prospective study. Am J Epidemiol 1994; 139: 11929.

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. MATERIALS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. AUTHOR CONTRIBUTIONS
  8. Acknowledgements
  9. REFERENCES
  10. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ACR_20464_sm_appendix.doc148KSupplementary Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.