Fig. 1

Multivariable logistic regression with baseline serum uric acid (UA) as predictor of the metabolic syndrome and its components after 7 years, stratified into normal-weight and overweight at baseline by body mass index (BMI). The odds ratio is per 59 μmol/L increase of UA. Bars represent odds ratio; grey bars the normal-weight subjects (BMI < 25 kg/m²) and black bars the overweight subjects (BMI ≥ 25 kg/m²). Whiskers represent 95 % confidence interval. The group includes the subjects without each component of MetS of interest or MetS (three or more components) at baseline. Covariates: baseline sex, age, systolic blood pressure, total cholesterol, current smoking, physical activity, HbA1c, eGFR, alcohol consumption, use of diuretics, and waist circumference. * = P for interaction with BMI-cut-off <0.05. Elevated blood pressure: blood pressure ≥130/85 mm Hg or treated for hypertension, elevated triglycerides: triglycerides ≥2.28 mmol/L if time since last meal <4 h and ≥1.7 mmol/L if time since last meal ≥4 h or use of lipid lowering drugs, low HDL = HDL < 1.03 mmol/L in men or <1.30 mmol/L in women, elevated fasting glucose = glucose ≥7.8 mmol/L if time since last meal <4 h and ≥5.6 mmol/L if time since last meal ≥4 h or treated for elevated glucose, central obesity = waist circumference ≥102 cm in men or ≥88 cm in women, MetS = three or more MetS components. Abbreviations: HDL high-density lipoprotein, BMI body mass index, MetS metabolic syndrome, UA uric acid, HBA1c hemoglobin A1c, eGFR estimated glomerular filtration rate