Table 1.

Epidemiology studies reporting positive association between hyperglycaemia or diabetes mellitus and osteoarthritis

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
Prospective cohort studies
Yoshimura 2012, Japan (ROAD study)20	CH	1690	Knee	X-ray	Mean BMI OA: 23.6±2.9 kg/m2 Non-OA: 22.4±3.2 kg/m2	Age, gender and components of MetS (include BMI≥25)	***	BMI* Age (∅)	**	DM associated with prevalence of knee OA adj OR=1.94 (95% CI 1.05 to 3.59) DM associated with knee OA progression adj OR=1.18 (95% CI 0.84 to 1.64)
Schett 2013, Italy7	CH	927	Hip and knee	Arthroplasty	Mean BMI DM: 27.0±3.9 kg/m2 Non-DM: 24.8±3.7 kg/m2	Age, sex, BMI and other variables	****	BMI (∅) Age (∅)	**	Significant increase risk of knee OA, comparing DM to non-DM adj HR=2.09 (95% CI 1.10 to 3.99)
Davies-Tuck 2012, Australia23	CH	165	Knee	MRI cartilage volume and bone marrow lesion	No data	Change in BMI over time	****	BMI (∅)	**	Fasting blood glucose was positively associated with incident bone marrow lesions in women, adj OR=5.76 (95% CI 1.06 to 31.21) but not in men, adj OR=0.11 (95% CI 0.01 to 1.79)
Cross-sectional studies
Wang 2013, China24	CS	1877	Knee and hand	X-ray and symptoms	No data	Age, sex, BMI, comorbidities	***	BMI (∅) Age (∅)	*	DM was associated with OA. Hand OA, OR=2.711 (95% CI 0.397 to 8.904) Knee OA, OR=2.247 (95% CI 0.746 to 12.370)
Siviero 2009, Italy Italian (Longitudinal study of Ageing)25	CS	1867	Shoulder	Physician diagnosis	Mean BMI: Definite OA: 27.3±4.8 kg/m2 Non-OA 26.6±4.1 kg/m2	Age, sex, BMI and others	***	BMI (∅) Age (∅)	***	Higher prevalence of DM in OA Shoulder OA: 9.4%
Haugen 2015, USA21	Cross-sectional data of prospective cohort	1348	Hand	X-ray and symptoms	Mean BMI Radiographic OA: 27.8±4.9 kg/m2 Symptomatic OA: 28.0±4.4 kg/m2 Non-OA: 27.4±4.5 kg/m2	No adjustment for prevalence data	****	BMI*	**	Higher prevalence of DM, comparing hand OA to no hand OA
Hart 1995, USA (The Chingford study)26	CS	979 Women only	Knee	X-ray and symptoms	No data	Age, BMI	****	BMI (∅) Age (∅)	****	Significant increase risk of knee OA in subjects with third tertile compared with first tertile of glucose level.
Eymard 2015, Multicentre, 18 countries (SEKOIA trial)27	Posthoc analysis of placebo arm of RCT	559	Knee	OA progression by annualised JSN	DM vs non-DM: No significant difference in joint space width	Age, sex, BMI, hypertension dyslipidemia	**	BMI (∅) Age (∅)	**	Significant higher annualised JSN in DM men only (p=0.027), but not women with DM.
Reid 2010, USA28	CS	6299	Not specified	ICD code	Not available	Age and gender in separate models. Adjusted for BMI in categories, hypertension, smoking	***	BMI (∅) Age*	*	Significant increased risk of OA in younger women, comparing DM vs non-DM No significant risk of OA in older women and men, comparing DM vs non-DM
Case-control studies
Rahman 2013, Canada (Canadian Community Health Survey)19	CC	81 634	Not specified	Structured questionnaire	Prevalence of Obesity (BMI ≥30 kg/m2) OA 20.5% Non-OA=14.3%	No adjustment for prevalence data	***	Age*	**	Higher prevalence of DM, comparing OA to non-OA.
Puenpatom 2009, USA (NHANES III data)43	CC	7714	Not specified	Record linkage, X-ray, self-report	No data	No adjustment for prevalence data	***	Nil	**	Higher prevalence of DM, comparing OA to non-OA
Cimmino 1990, Italy44	CC	1246	Finger, hip, knee, spine, diffuse	X-ray	Prevalence of Obese† OA: 18% Non-OA: 16.4%	Nil	****	Nil	**	Significant higher blood glucose in OA vs non-OA (both men, women, both sex)
Karvonen-Gutierrez 2012, USA (NHANES III data)22	CC	1066	Knee	X-ray	Mean BMI OA: 29.4 kg/m2 Non-OA: 26.0 kg/m2 (p<0.0001).	Nil	***	BMI (∅)	***	Significant association between HOMA-IR and knee OA. Non-obese men, unadj OR=1.18 (95% CI 1.15 to 1.22) Obese men, unadj OR=1.34 (95% CI 1.27 to 1.42) Non-obese women, unadj OR=1.04 (95% CI 1.01 to 1.07) Obese women, unadj OR=0.88 (95% CI 0.86, 0.89)
Maddah 2015, Iran45	CC	625	Knee	X-ray	Mean BMI for men: OA: 23.3±3.1 kg/m2 Non-OA: 23.5±3.9 kg/m2 Mean BMI for women: OA: 22.8±4.3 kg/m2 Non-OA: 25.1±4.6 kg/m2 (p<0.001)	Nil	***	Nil	**	Higher prevalence of DM among women with knee OA compared with non-OA Not significant for men
Nieves-Plaza 2013, Puerto Rico8	CC	202	Hand and knee	ACR criteria	Prevalence of Obesity (BMI ≥30 kg/m2) OA: 56.6% non-OA: 36.5%	Age, sex, obesity (BMI ≥30 kg/m2, and other variables	*	BMI (∅) Age (∅)	**	Significant increased risk of hand/ knee OA, comparing DM to non-DM
Waine 1961, USA6	CC	60	Knee	X-ray	Prevalence of Obese‡ DM: 36.7% Non-DM: 16.7%	Matched age in case/ control BMI not adjusted	*	Age*	*	Higher X-ray scores in DM compared with non-DM subjects.

*§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.

†Obese defined as weight exceeding ideal weight by 10 kg.

‡Obese defined as exceeding 20% of ideal body weight.

ACR, American college of rheumatology; adj, adjusted; BMI, body mass index; CC, case-control study; CH, prospective cohort study; CS, cross-sectional study; DM, diabetes mellitus; HOMA-IR, homeostatic model assessment-insulin resistance; ICD, international classification of diseases; JSN, joint space narrowing; KL, Kellgren-Lawrence grade; metS, metabolic syndrome; n, sample size; OA, osteoarthritis; RCT, randomised controlled trial; unadj, unadjusted; X-ray, radiographic.