Figure 3.

Associations of types of dairy consumption (servings/d) with waist circumference and the waist-to-hip ratio among over 12,000 adults in the Fenland Study, UK. Forest plots represent regression coefficients with their 95% CIs adjusted for age (years), sex, test-site (Cambridge, Ely, Wisbech), ethnicity (white, non-white), total energy intake (kcal/d), other dairy types, educational level (low, medium, high), age when full-time education finished (years), socio-economic status based on occupation (low, medium, high), income (<£20,000, £20,000-40,000, >£40,000), marital status (single, married, widowed/separated), smoking status (never, former, current smoker), pack-years of smoking, energy expenditure due to physical activity (kJ/kg per d), lipid-lowering medication (yes/no), anti-hypertensive medication (yes/no), hormone-replacement therapy (yes/no in women), intakes (g/d) of fruit, vegetables, potatoes, legumes, processed cereals, whole-grain cereals, poultry and eggs, red meat, processed meat, fish, sauces, margarine, nuts, sweet snacks, sugar-sweetened beverages, artificially sweetened beverages, fruit juice, regular coffee, decaffeinated coffee, tea and alcoholic beverages, plasma vitamin C levels (μmol/l), dietary supplement use (yes/no), and BMI (kg/m²). Statistically significant associations after false discovery rate corrections are marked with an asterisk. See categorisation of dairy types in Table 1.