Figure - PMC

Skip to main content

View full-text article in PMC

. Author manuscript; available in PMC: 2014 May 1.

Published in final edited form as: Biofactors. 2013 Jan 28;39(3):242–258. doi: 10.1002/biof.1077

New onset disease (Incidence) by whole grain intake (in grams dry-weight per day) (see Table 1).

a. Whole grain intake and new onset T2D (see Table 1). Common slope is −0.039 ±0.005 (slope units: 1 case/1000 person years (incidence) per 1 gram of whole grain (dietary intake)). That is, the incidence of T2D drops by 1.57 cases/1000 person years (95% CI 1.14–1.99) for an increase of 40 grams of whole grains per day.

(Analysis of covariance regression model: (incidence= whole_grain [continuous variable] + study [class variable] + whole_grain × study [interaction effect]). For BWHS the data are more accurately described by an inverse function (r²=0.99) than linear negative (r²=0.62, p=0.11) with the effect of whole grains appearing to flattening out at higher whole grain consumption).

b. New onset of CHD by quintiles of whole_grain intake (see Table 1) (note the logarithmic scale). The CHD incidence drops 26% (95% CI 19–33%) for a 20 gram daily increase in whole grain intake (linear portions of regressions). The CHD incidence drops 36% (95% CI 27–45%) for a 30 gram daily increase in whole grain intake, but this may only apply to those populations where the slopes plateau at ≥30 grams per day (For ARIC, quintile person-years are calculated from total person-years).

(Analysis of covariance regression model with spline, as graphed: (Log_incidence=whole_grain [continuous variable] + study [class variable] + spline × (whole_grain - elbow) [spline effect variable (where spline=0 below elbow and 1 above elbow (in grams) and elbow=point of the bend)]). For the linear parts of Figure 4b; Analysis of covariance regression model: (Log_incidence=whole_grain [continuous variable] + study [class variable] + whole_grain × study [interaction effect]). For Log_Incidence v. whole_grain, p<0.0001, for interaction effect p=0.70 i.e. slopes are not different and regression solution obtained ignoring any interaction effect therefore).

c. New onset Hypertension (Incidence) by quintiles of whole grain intake (see Table 1). The slopes are significantly different (p<0.006). Ignoring this difference, the common slope by analysis of covariance is= −0.085 (p<0.007) (for reference, the risk ratio for women with multivariate adjustment was 0.89, p=0.007 in the original publication(7)). For men (10), in Figure 4c the slope= −0.134 ± 0.019 (p=0.006). A 40 gram daily increase of whole grains reduces the hypertension incidence by 5.4 cases/1000 person-years (95% CI 3.8–6.9), a drop of 20% (95% CI 14–26%).

New onset disease (Incidence) by whole grain intake (in grams dry-weight per day) (see Table 1).

a. Whole grain intake and new onset T2D (see Table 1). Common slope is −0.039 ±0.005 (slope units: 1 case/1000 person years (incidence) per 1 gram of whole grain (dietary intake)). That is, the incidence of T2D drops by 1.57 cases/1000 person years (95% CI 1.14–1.99) for an increase of 40 grams of whole grains per day.

(Analysis of covariance regression model: (incidence= whole_grain [continuous variable] + study [class variable] + whole_grain × study [interaction effect]). For BWHS the data are more accurately described by an inverse function (r²=0.99) than linear negative (r²=0.62, p=0.11) with the effect of whole grains appearing to flattening out at higher whole grain consumption).

b. New onset of CHD by quintiles of whole_grain intake (see Table 1) (note the logarithmic scale). The CHD incidence drops 26% (95% CI 19–33%) for a 20 gram daily increase in whole grain intake (linear portions of regressions). The CHD incidence drops 36% (95% CI 27–45%) for a 30 gram daily increase in whole grain intake, but this may only apply to those populations where the slopes plateau at ≥30 grams per day (For ARIC, quintile person-years are calculated from total person-years).

(Analysis of covariance regression model with spline, as graphed: (Log_incidence=whole_grain [continuous variable] + study [class variable] + spline × (whole_grain - elbow) [spline effect variable (where spline=0 below elbow and 1 above elbow (in grams) and elbow=point of the bend)]). For the linear parts of Figure 4b; Analysis of covariance regression model: (Log_incidence=whole_grain [continuous variable] + study [class variable] + whole_grain × study [interaction effect]). For Log_Incidence v. whole_grain, p<0.0001, for interaction effect p=0.70 i.e. slopes are not different and regression solution obtained ignoring any interaction effect therefore).

c. New onset Hypertension (Incidence) by quintiles of whole grain intake (see Table 1). The slopes are significantly different (p<0.006). Ignoring this difference, the common slope by analysis of covariance is= −0.085 (p<0.007) (for reference, the risk ratio for women with multivariate adjustment was 0.89, p=0.007 in the original publication(7)). For men (10), in Figure 4c the slope= −0.134 ± 0.019 (p=0.006). A 40 gram daily increase of whole grains reduces the hypertension incidence by 5.4 cases/1000 person-years (95% CI 3.8–6.9), a drop of 20% (95% CI 14–26%).

HHS Vulnerability Disclosure