Author /year /country (study name) | Study design | N | Site of OA | Definition of OA | Comparison of obesity at baseline | Adjustment of obesity/ age | Quality indicators | Main results | ||
Selection | Comparability (BMI and age)*§ | Outcome | ||||||||
Prospectivecohortstudies | ||||||||||
Leung 2017, Singapore (Singapore Chinese Health Study)9 | CH | 63 257 | Knee | Arthroplasty | Mean BMI:
| Age, sex, BMI and other variables | **** | BMI* Age (∅) | *** | No significant risk of knee OA, comparing DM to non-DM.
|
Monira Hussain 2014, Australia (Melbourne Collaborative Prospective cohort Study)29 | CH | 13 753 | Hip and knee | Arthroplasty | Mean BMI
| Age, sex, BMI and other variables | *** | BMI (∅) Age (∅) | ** | No significant risk of hip/knee OA, comparing DM to non-DM.
|
Engstrӧm 2009, Sweden (Malmo Diet and Cancer study)30 | CH | 5171 | Hip, knee | Arthroplasty | Mean BMI
| Age, sex, BMI and other variables | *** | BMI (∅) Age (∅) | *** | No significant risk of hip and knee OA, comparing DM to non-DM. Hip OA: adj RR=0.7 (95% CI 0.4 to 1.2) Knee OA: adj RR=1.4 (95% CI 0.9 to 2.4) |
Frey 1996, USA31 | CH | 1514 | Not specified | Self -reported | Mean BMI men.
Mean BMI women:
| Age, weight, BMI, max life time weight | ** | BMI* Age (∅) | ** | No significant difference in glucose levels or prevalence of DM, compare OA to non-OA. IGT and DM was not associated with a history of OA |
Driban 2017, USA OAI dataset32 | CH | 162 | Knee OA progression | X-ray | Mean BMI
| Age, and BMI | **** | BMI (∅) Age (∅) | ** | Glucose, glycated serum protein and C reactive protein were not significantly associated with incident or accelerated knee OA. |
Cross-sectionalstudies | ||||||||||
Anderson 1988 USA (HANES I)36 | CS | 5193 | Knee | X-ray | Not available | age | **** | Age* | nil | No significant risk of knee OA, comparing DM to non-DM Age adj OR, men=2.14 (95% CI 0.92 to 4.96) Age adj OR women=1.02 (95% CI 0.52 to 2.00) Null association in multi-variable regression model (data not shown) |
Dahaghin 2007, Netherlands (Rotterdam study)46 | CS | 3585 | Hand | X-day | Significant association between Hand OA and BMI >27.4 kg/m2, OR=1.4 (95% CI 1.2 to 1.7) | Age, sex, BMI and other variables | *** | BMI (∅) Age (∅) | ** | No significant risk for hand OA, comparing DM to non-DM adj OR=1.2 (95% CI 0.9 to 1.6) |
Sturmer 2001 Germany (Ulm OA study)38 | CS | 809 | Bilateral OA or Generalised OA | X-ray and physical examination | Prevalence of obesity (BMI ≥30 kg/m2)
| Age, sex, BMI and other variables | ** | BMI (∅) Age (∅) | *** | For patients undergoing hip/knee arthroplasty, DM was not associated with:
|
Martin 1997, USA (Baltimore longitudinal study of ageing)47 | CS | 739 | Knee | X-ray | Mean BMI for men
Mean BMI for women
(p=0.0007) | Age and BMI | *** | BMI (∅) Age (∅) | *** | No significant risk of knee OA, comparing DM to non-DM.
|
Bagge 1991, Sweden (70 year old people in Goteborg)48 | CS data from prospective cohort | 340 | Knee and hand | X-ray | Significant association between:
| BMI | *** | BMI (∅) Age* | *** | No significant risk of knee OA, comparing DM and non-DM. Women: RR=1.44 (95% CI 0.54 to 3.0); Men: adj RR=1.08 (95% CI 0.51 to 2.28); Both sex: adj RR=1.22 (95% CI 0.72 to 2.07). Null association for Hand OA (data not shown) |
Case-controlstudies | ||||||||||
Nielen 2016, UK12 | CC | Hip: 99 682 Knee: 89 536 | Hip and knee | Arthroplasty | Mean BMI for hip OA
Mean BMI for knee OA
| Matched age and sex. Adjusted for: BMI and smoking | ** | BMI* Age (∅) | ** | No significant risk of hip/knee OA, comparing DM to non-DM.
Current antidiabetic drug use and increasing HbA1c associated with lower risk of hip/ knee arthroplasty |
Han 2013, Korea (Korea National Health and Nutrition Examination Survey)34 | CC | 2234 | Knee | Self-reported +symptoms | Mean BMI
| Age, height exercise, alcohol intake and smoking | ** | BMI* Age (∅) | ** | No significant difference in fasting blood glucose, compare OA vs non-OA No significant risk of knee OA, comparing DM to non-DM
|
Al-Jarallah 2016, Kuwait13 | CC | 311 | Knee | X-ray | Prevalence of BMI≥30 mg/kg2:
| Age, sex, BMI | *** | BMI (∅) Age (∅) | ** | No significant association with presence of osteophytes, comparing DM and non-DM, adj OR=0.098 (95% CI 0.042 to 0.230) |
Magnusson 2017, Norway35 | CC | 165 | Hand | X-ray | Mean BMI
| Age, sex, waist circumference and education | *** | BMI* Age* | ** | No significant risk of radiographic OA or erosive hand OA, comparing type 1 DM to non-DM
Among subjects with long-term type 1 DM, hyperglycaemic burden (mean time weighted HbA1c) not associated with radiographic hand OA and erosive hand OA |
Horn 1992, USA33 | CC | 73 | Knee | X-ray and symptoms | Mean weight
| Matched age and BMI in case/ control | * | BMI* Age* | ** | Significant lower prevalence of osteophytosis, comparing DM to non-DM women:
P=0.0044 |
*§Comparability scoring is given as full star (*) to comparable BMI and age at baseline; and half star (denoted by (∅)) if BMI and age was adjusted with statistical modelling.
adj, adjusted; BMI, body mass index; CC, case-control study; CH, prospective cohort study; CI, confidence intervals; CS, cross-sectional study; DM, diabetes mellitus; HbA1c, glycated haemoglobin; HOMA-IR, homeostasis model assessment-estimated insulin resistance; HR, HR ratio; HANES I, the United States first national Health and Nutrition Examination Survey of 1971–1975 (HANES I); IGT, impaired glucose tolerance; KOA, knee osteoarthritis; MetS, metabolic syndrome; n, sample size; OA, osteoarthritis; OAI, Osteoarthritis initiatives; OR, odd ratio; RR, relative risk; T2, T2-weighted MRI scan; X-ray, radiographic.