Serum triglycerides in the prediction of coronary artery disease (an Italian experience).

An occupational group living in Rome and composed of 3,007 men aged 46 to 65 years who were free from previous major coronary events was screened for a number of coronary risk factors and then followed up for 10 years. In all, 107 coronary deaths occurred in 10 years. There was a positive relation between coronary death rate and increasing levels of triglycerides, but the difference between the extreme quintile classes was not any more significant after adjustment for cholesterol levels. A cross-classification involving low-density lipoprotein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol and triglyceride levels showed a higher coronary artery disease (CAD) mortality in subjects with a higher LDL cholesterol/HDL cholesterol ratio and higher triglycerides. However, the excess risk in this subgroup was largely explained by the mean levels of total cholesterol. The multivariate proportional-hazards Cox model with coronary deaths as the end point, and age and triglycerides as predictors, produced a significant coefficient for triglycerides that became nonsignificant when other lipids alone or in combination (total, HDL and LDL cholesterol and some of their ratio) were fed into the model as further covariates. It is concluded that there is no evidence of an independent role of triglycerides in the prediction of coronary deaths.