Abstract: Objectives: To compare the abilities of four lipid predictors to predict the population at risk of atherothrombotic disease (ATD).Methods:The lipid predictors are: LDL, CT, CT:HDL, and (LDL-HDL)/LDL.The author's general and ATD population databases demonstrate the frequencies and severity of each lipid predictor.Results: If each predictor is divided into sextiles, then as each predictor passes from lowest to highest sextile, the percentage of ATD patients per sextile increases progressively; however, the number of people in the general population per predictor sextile decreases progressively for LDL, CT, and CT:HDL-only for (LDL-HDL0/LDL does the number of people increase.The sextile of each lipid predictor in the ATD population in which the most patients are found is in the mid to lower sextile for LDL, CT, and CT:HDL.Only for (LDL-HDL)/LDL are the most patients found at the high end of the sextiles.At any level of LDL, CT, CT:HDL, knowledge of (LDL-HDL0/LDL allows further risk stratifi cation in terms of age of ATD onset.A lower (LDL-HDL)/ LDL is associated with an older age of ATD onset; a higher (LDL-HDL)/LDL, with a younger age.The reverse is not uniformly true.Conclusion: (LDL-HDL)/LDL is the best lipid predictor because, unlike LDL, CT, and CT:HDL, it predicts ATD risk progressively better in more patients as predictor sextiles range from lowest to highest and it has most ATD patients in its highest sextiles.Also, (LDL-HDL)/LDL can further risk-stratify any level of LDL, CT, and CT:HDL.