Abstract P125: Reverse Causality of the Relationship Between Dietary Cholesterol and Serum Low Density Lipoprotein Cholesterol Associated With Education and Employment: the INTERLIPID Study

Abstract

Introduction: The upper limit of recommended dietary cholesterol has been abolished recently from guidelines in Japan and US even though high intake is known to increase serum cholesterol levels. Reverse causality has been involved, reflecting availability of widespread information with consequent lifestyle modifications among knowledgeable people. People with higher serum cholesterol are more likely to modify diet than those with lower serum cholesterol, especially among more educated; if enough people do this, dietary cholesterol will not relate to serum cholesterol or will relate inversely. Hypothesis: We assessed the hypothesis that lifestyle modifications separately by those with high education level and employment status reverse the positive relationship between dietary cholesterol and serum low density lipoprotein cholesterol (LDL-C); therefore, the positive relationship between dietary cholesterol and LDL-C would prevail after adjustment for education level and employment status. Methods: A population-based, random sample, cross-sectional study (INTERLIPID) which was an ancillary study of the International Study of Macronutrients and Blood Pressure (INTERMAP), conducted a survey between 1996 and 1999. Among 1,145 Japanese individuals ages 40-59 years, 106 were excluded because of special diets, use of lipid lowering drugs, hormone replacement, and missing data, leaving 1,039 individuals (533 men and 506 women). Dietary cholesterol was assessed from four 24-h dietary recalls and LDL-C was measured by enzymatic methods on an auto-analyzer. A standard questionnaire inquired about years of education and employment status. Results: Linear regression analysis revealed an inverse association between dietary cholesterol and LDL-C in men (coefficient, -0.04 [95%CI, -0.08 to -0.01]), no significant association when adjusted for education and its interaction term with dietary cholesterol in men (coefficient, 0.05 [95%CI, -0.03 to 0.14]), and a significant positive relation when adjusted for employment and its interaction term in men (coefficient, 0.13 [95%CI, 0.01 to 0.26]). With adjustment for education and employment and their interactions, a significant positive relation was recorded again in men (coefficient, 0.16 [95%CI, 0.03 to 0.29]). These relationships were not observed in women. Conclusions: In Japanese men, education, employment, and their interaction terms with dietary cholesterol reversed an observed inverse relationship between dietary cholesterol and LDL-C. A positive relation prevailed between dietary cholesterol and LDL-C after adjustment for education level and employment status.