Table 3

Summary of risk estimates of the association between diabetes or hyperglycaemia and osteoarthritis

No of studiesOR (95% CI), p valueI2 (%)P for heterogeneityEgger’s test for publication bias (p value)
Overall221.14 (0.91 to 1.03), p=0.1074.2<0.010.02
Comparability
 Studies with any adjustment of BMI7–9 12 13 20 23–26 28–31 34 35 38 46–48201.14 (0.97 to 1.34), p=0.1074.3<0.010.01
 Studies with comparable baseline BMI between groups9 20 31 33–3560.94 (0.68 to 1.29), p=0.6872.6<0.010.23
Study design
 Prospective cohort7 9 20 23 29–3171.08 (0.80 to 1.44), p=0.6376.0<0.010.14
 Case-control8 12 13 33–3560.82 (0.49 to 1.38), p=0.4585.7<0.010.89
 Cross-sectional24–26 28 36 38 46–4891.29 (1.10 to 1.51), p<0.0130.00.100.68
Gender
 Men9 23 28 31 34 36 47 4881.00 (0.76 to 1.31), p=0.9949.40.030.22
 Women9 23 26 28 31 34 36 47 4891.24 (0.92 to 1.67), p=0.1675.6<0.01<0.01
Site of OA
 Hand OA24 31 35 4641.26 (0.77 to 2.05), p=0.3669.00.010.66
 Hip OA12 29–3140.90 (0.74 to 1.09), p=0.2752.50.080.88
 Knee OA9 12 13 20 23 24 26 29–31 33 34 36 47 48150.97 (0.79 to 1.18), p=0.7572.1<0.010.27
  • BMI, body mass index; CI, confidence intervals; OA, osteoarthritis; OR, odd ratio.