Figure 1 Correlation of DAS28 with the number of metabolic syndrome (MetS) components present. Patients were grouped depending on the number of MetS components into three groups: those with no MetS components present (n = 21), those with 1–2 MetS components present (n = 91) and those with ⩾3 components (n = 88). A significant linear correlation of DAS28 with the number of MetS components present was found after controlling for age and sex effects (β correlation coefficient 0.237, p = 0.001) or after controlling for demographics (age, sex), rheumatoid arthritis treatments (MTX, steroids, biologicals) and behavioural characteristics (smoke and BMI; β coefficient = 0.231, p = 0.003).