Personal and familial factors in cholesterolemia: criteria for selection of a reference population.

We studied factors affecting cholesterolemia in a presumably healthy population of 23 300 subjects, ages 20 to 59 years. Multiple-contingency tables were established, to quantify the information obtained by means of a factorial correspondence analysis. The relations between values for cholesterol and those for triglycerides, uric acid, and overweight were constant for all ages and both sexes. Cholesterolemia paralleled alcohol and tobacco consumption and arterial blood pressure. Use of oral contraceptives did not influence cholesterol values. We classified variation factors according to the statistical method of segmentation, by sex and age. This classification permitted us to establish criteria for excluding subjects from a population sampled with the aim of producing reference values for plasma cholesterol. The values of cholesterol for the 50th percentile were 5.40 mmol/L for men 20 to 39 years old and 5.79 mmol/L for those older than 40. For women, the 50th percentile values increased from 5.34 mmol/L between ages 20 to 39 to 5.92 mmol/L for ages 40 to 49 and to 6.51 mmol/L after 50.