OBJECTIVE: In patients with type 2 diabetes, a normal HDL cholesterol level does not rule out that LDL particles may be small. Although techniques for analyzing LDL subfractions are not likely to be used in clinical practice, a prediction of LDL size based on a regular lipid profile may be useful for assessment of cardiovascular risk. RESEARCH DESIGN AND METHODS: Sixty patients with type 2 diabetes with acceptable glycemic control and an HDL cholesterol level > or = 1 mmol/l were recruited after cessation of lipid-altering treatments. LDL size was determined by 2-20% PAGE; patients having small LDL (n = 30) were compared with those having intermediate or large LDL (n = 30). RESULTS: Clinical characteristics, pharmacological therapies, lifestyle, and prevalence of diabetes-related complications were similar in both patient groups. LDL size correlated negatively with plasma triglycerides (TGs) (R2 = 0.52) and positively with HDL cholesterol (R2 = 0.14). However, an inverse correlation between the TG-to-HDL cholesterol molar ratio and LDL size was even stronger (R2 = 0.59). The ratio was > 1.33 in 90% of the patients with small LDL particles (95% CI 79.3-100) and 16.5% of those with larger LDL particles. A cutoff point of 1.33 for the TG-to-HDL cholesterol ratio distinguishes between patients having small LDL values better than TG cutoff of 1.70 and 1.45 mmol/l. CONCLUSIONS: The TG-to-HDL cholesterol ratio may be related to the processes involved in LDL size pathophysiology and relevant with regard to the risk of clinical vascular disease. It may be suitable for the selection of patients needing an earlier and aggressive treatment of lipid abnormalities.