Sources and magnitudes of relative risks for the effects of continuous dietary risk factors on disease-specific mortality.

aFor these risk factor�Cdisease pairs, RRs in the source were reported for all ages combined. We used median age at event and the age pattern of excess risk for serum total cholesterol and IHD to estimate RRs for each age category.bThe interventions studies replaced dietary SFA with PUFA, hence the RRs measure the effect of replacement. Effects of replacing PUFA for other macronutrients have not been evaluated in randomized interventions studies. However, evidence from cohort studies suggests that replacement of PUFA for carbohydrates, but not carbohydrates for SFA, would produce similar benefits [78], indicating that the measured benefits are due to PUFA.