233 - Exploring lipoprotein patterns in Di@bet.es study (ID 1022)
Abstract
Background and Aims
The Di@bet.es Study was a Spanish cross-sectional, population-based study conducted in 2009-10 to evaluate Diabetes Mellitus (DM) prevalence. The aim of the present study was to define lipoprotein patterns according to 2D-diffusion-ordered 1H-Nuclear Magnetic Resonance (NMR) spectroscopy data.
Methods
Study group included 4826 subjects. The Liposcale Test was used to quantify the lipid content, particle number and size of lipoprotein subclasses from serum samples. We performed exploratory multivariate analysis (unsupervised k-means clustering) to identify different lipoprotein patterns among major traditional lipid profiles.
Results
The variables included in the models for pattern recognition algorithm were total cholesterol (TC), triglycerides (TG), prevalence of small particles and lipid composition of each lipoprotein fraction (VLDL, IDL, LDL and HDL). Most of normolipidemic individuals (30%, TC<200 mg/dl; TG<150 mg/dl) presented a low-risk lipoprotein profile, although there was a group (10%, 57%men, age 50±18, TG 110±28, LDL-C 116±19, HDL-C 52 ± 11mg/dl, 36% obesity, 14%DM) that presented HDL particles enriched in triglycerides. Hypercholesterolemic individuals (50%, TC>200mg/dl, TG<150 mg/dl) showed increased cholesterol composition in VLDL particles and large LDL particles. Combined hyperlipidaemia individuals (20%, TC>200 mg/dl; TG>150 mg/dl) could be reclassified on the bases of markedly elevated triglycerides in LDL and HDL fractions. There was a subgroup with severe hyperlipidaemia (0,2%, 88%men, age 44±11) which presented small HDL particles as well as IDL particles depleted of triglycerides and enriched in cholesterol.
Conclusions
NMR derived lipoprotein pattern recognition beyond standard lipid values allows a broad analysis of lipoprotein disturbances, a better stratification of patients and thus a more accurate clinical assessment.
