Relative risk (95% CI) for anemia at follow-up derived from Poisson regression according to quartiles of baseline riboflavin intake among Chinese adults by anemic status at baseline in Jiangsu Nutrition Study.^a

aValues are prevalence rate ratio (95% CI) from Poisson regression. bmean riboflavin intake with quartile. cModel 1 adjusted for age (years, as continuous), sex, and energy intake. dModel 2 adjusted for age, sex, smoking (0, 1–19, ≥20 cigarettes/day), alcohol drinking (g/day), education (low, medium, high), and occupation (manual/non-manual), overweight (BMI≥24 kg/m2, yes/no), hypertension (yes/no), intake of energy (kcal/day), iron (mg/day), and vitamin C (mg/day) (as continuous variables), and baseline dietary patterns. eModel 3 additional adjusted for dietary patterns (continuous) at follow-up.