Association of Whole Grain Intake with Diabetes and Subclinical CVD in the Multiethnic Study of Atherosclerosis: A Sensitivity Analysis

Abstract

Background: A dietary pattern that includes whole grains is widely believed to provide numerous health benefits, such as a reduction in chronic disease risk factors like inflammation, insulin resistance, and subclinical markers of cardiovascular disease. A paper in 2007 by Lutsey et al. looked at the relationship between whole grains and CVD biomarkers. The results of this study found an inverse association between whole grain intake and diabetes risk but were unable to find significant results with CVD biomarkers. In the time since the publication of this paper, MESA corrected an error in how dark bread was counted in total whole grain intake. The purpose of this study was to replicate the analysis in Lutsey et al. using the corrected whole grain intake. Methods: Whole grain intake was assessed using a 127-item food frequency questionnaire collected between 2000 and 2002 during MESA exam 1. We applied the same exclusion criteria as Lutsey, et al (n=5493) and compared them against the five original outcomes; obesity, insulin resistance, inflammation, newly diagnosed diabetes, and subclinical CVD. Results: In our analysis, the whole grain data modification greatly altered mean intake and shifted subjects around within the five study quintiles (mean intake of Q1=0.04g; Q2=0.32g; Q3=0.65g; Q4=1.03g; Q5=1.79g). However, our replication of the paper did not change key outcomes of Lutsey et al. Conclusion: Our re-analysis supports the association between whole grain intake and diabetes and no relationship was found with subclinical CVD in the MESA cohort. Additionally, this sensitivity analysis demonstrates the limitations of using FFQs and self-reported dietary records to inform national and international guidelines.