AIMS: Low-density lipoprotein cholesterol (LDL-C) levels are often fairly normal in Type 2 diabetes mellitus (DM). We anticipated that a parabolic relation between plasma triglycerides and LDL-C, as previously demonstrated in familial combined hyperlipidaemia (FCHL), might account for this phenomenon. METHODS: Our hypothesis was tested in 1343 subjects derived from the general population who were studied on two occasions 6 years apart (the Hoorn study). Three groups were constructed depending on plasma triglycerides: group A (individuals with both measurements below 1.5 mmol/l), group B (one measurement below and one above 1.5 mmol/l) and group C (both measurements above 1.5 mmol/l). Diabetes status was ascertained by an oral glucose tolerance test. RESULTS: In a mixed linear model, a significant, positive relation between triglycerides and LDL-C was observed for males in group A (beta(a) = 0.5, P < 0.001) and group B (beta(b) = 0.2, P < 0.001), whereas a significant negative relation was found for males in group C (beta(c) = -0.2, P = 0.003). The regression slopes did not differ between diabetic and non-diabetic subjects. Similar results were obtained for women, with the exception that the relation was not significantly negative in group C (beta(c) = -0.1, P = 0.4). CONCLUSION: Plasma triglycerides and LDL-C are related in a parabolic fashion, not only in FCHL, but also in the general population and Type 2 DM. These findings aid our interpretation of typical dyslipidaemia and the effects of treatment that are frequently observed in hypertriglyceridaemic states.