What is the best predictor of the atherogenic LDL subclass phenotype 'pattern B' in patients with type 2 diabetes mellitus?

Background: The atherogenic lipoprotein phenotype 'pattern B' comprises a predominance of small-dense low-density lipoprotein (sdLDL). Gradient gel electrophoresis (GGE) is considered a 'gold standard' method for identifying this phenotype, but is impractical for routine laboratory use. The low-density lipoprotein cholesterol:apolipoprotein-B (LDL-C:Apo-B) ratio has been advocated as a surrogate marker for sdLDL and a direct assay for sdLDL has recently become available. We compared the sdLDL assay and LDL-C:Apo-B with more established lipid parameters to predict the presence of 'pattern B' phenotype. Method: Blood was collected from 97 fasted subjects on three separate occasions. Total cholesterol, triglyceride, Apo-B and sdLDL were measured; LDL- and HDL-cholesterol were determined after ultracentrifugation. The predominant LDL particle size and phenotype were assigned by GGE. Results: 'Pattern B' phenotype was identified in 36% of samples. Peak particle size showed a positive correlation with HDL-cholesterol and a negative correlation with triglyceride and Apo-B. Receiver operating curve (ROC) analysis showed triglyceride:HDL-C ratio and triglyceride alone to be the best predictors of 'pattern B' phenotype, with area under the curve (AUC) being 0.87 and 0.84, respectively. AUCs for sdLDL (0.74) and LDL-C:Apo-B (0.71) were significantly lower (P < 0.05). A high sdLDL concentration had the greatest specificity (95%) and positive predictive value (74%) for 'pattern B' phenotype, but low sensitivity (43%). Conclusion: Direct measurement of sdLDL provided the most specific predictor of 'pattern B' phenotype, whereas triglyceride:HDL-C ratio or triglycerides alone, parameters readily available in most laboratories, were the best predictors by ROC analysis.