|
1000
|
Abstract/Summary
|
-
BACKGROUND: Studies have consistently shown that risk of type 2 diabetes (T2D) is positively associated with dietary haem iron intake and inversely associated with dietary magnesium intake in a dose-response way. However, interaction effects of these two clinically important nutrients on T2DM risk in a prospective setting is unknown. OBJECTIVE: To determine the five-year risk of developing impaired fasting glucose (IFG) associated with dietary magnesium-to-iron intake ratios (Mg/Fe), including Mg/total Fe, Mg/haem-Fe and Mg/non-haem Fe. DESIGN: A cohort study of 1056 participants recruited into the Jiangsu Nutrition Study (JIN) from 2002 to 2007, aged at least 20 years and without known diabetes and IFG at baseline were followed up for five years. Dietary magnesium and iron intake at baseline was assessed by 3-day weighed food records. Fasting plasma glucose was measured both at baseline and follow up. Logistic regression models were performed to determine the associations between quartiles (using bottom quartiles as referent categories) of magnesium to iron (including total Fe, haem-Fe and non-haem Fe) ratio and the risk of IFG (>5.6 mmol/L) adjusted for covariates: age, gender, body mass index (BMI), hypertension, serum ferritin, haemoglobin and family history of diabetes. RESULTS: The mean (SD) intake of total Fe and magnesium was 25.0 (9.2) mg/d and 323 (125) mg/d. The incidence of IFG during 5-year follow up was 11.8%. Inverse associations were found between quartiles (Q) of Mg/haem-Fe and the risk of IFG in the fully adjusted model: odds ratios (OR) were 1.00, 0.59 (95%CI 0.35, 0.98), 0.49 (95%CI 0.28, 0.84), and 0.28 (95% CI 0.14, 0.55) (Q4), respectively. Weaker associations were found for quartiles of Mg/total Fe and no association was found between Mg/non-haem Fe and IFG risk. CONCLUSIONS: Low Mg/haem-Fe ratio is an independent risk factor for developing IFG in Chinese adults. Future research to determine the added predictive value of assessment of low dietary Mg/haem-Fe ratio beyond current T2D risk models in specific populations is justified.
|
