Liu K, Stamler J, Trevisan M, Moss D
Arteriosclerosis. 1982 May-Jun;2(3):221-7. doi: 10.1161/01.atv.2.3.221.
Univariate analyses with data for the years 1954-1969 from 20 economically advanced countries showed that a combined dietary lipid score based on per capita consumption of saturated fat, cholesterol, and polyunsaturated fat (equation of Keys, et al.) and the consumption of refined and processed sugars were both highly correlated with age-standardized coronary heart disease (CHD) mortality rates for the years 1969-1973 for both men and women aged 35-74 years (r values of 0.54 to 0.72). Fiber intake, estimated as the sum of calories available from vegetables, fruits, grains, and legumes, yielded a significant inverse correlation with CHD mortality rates (r = -0.49 to -0.68). These three dietary variables are highly intercorrelated. With bivariate analyses (2 X 2 cross-classification and analysis of variance), lipid score, but neither that of sucrose or fiber, was consistently and significantly related to CHD mortality. However, with both higher and lower lipid scores, CHD mortality rates tended to be slightly (nonsignificantly) higher with higher sugar values and with lower fiber values.
