ABC | Volume 110, Nº4, April 2018

Original Article Cartolano et al LAP and lipoprotein size Arq Bras Cardiol. 2018; 110(4):339-347 The distribution of HDL and LDL subfractions was determined using the Lipoprint supplier system based on nondenaturing polyacrylamide gel. The LDL1 and LDL2 sub‑fractions were classified as large LDL, and sub-fractions from LDL3 to LDL7 were classified as smaller and denser particles. The LDL size (nm) was determined and from that, phenotype A (> 25.6 nm, large and less dense LDL) and non-A (≤ 25.6 nm, small dense LDL) pattern were calculated. For HDL particle size, ten sub-fractions were identified, which were classified as large (HDL1 to HDL3), intermediate (HDL4 to HDL7), and small (HDL8 to HDL10) particles. All analyses were conducted in duplicate and intra- (1‑5.8%) and inter- (0.5-15%) assay coefficients of variance were calculated. Lipid Accumulation Product (LAP) LAP was calculated using different formulae for women (WC [cm] -58) × (TG [mmol/L] ) and men (WC [cm] -65) × (TG [mmol/L] ), which include the minimum sex-specific WC values. 8 Statistical analysis Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS®; v. 20.0) software package. Two-sided P values < 0.05 were considered statistically significant. The Kolmogorov–Smirnov test (p > 0.05) was applied to assess normality of data. Normally distributed continuous variables are presented as mean values and standard deviations (SD), whereas non-normally distributed data are presented as median and 25th and 75th percentiles. Categorical variables are presented as absolute values (n) and percentages (%). Groups were compared using the unpaired Student’s t-test for normally distributed data. Non-normally distributed data were analyzed using non-parametric Mann–Whitney U tests. Categorical variables were compared using the Pearson chi-square or Fisher’s exact test. Subjects were divided into tertiles (T) of the LAP index: T1 ≤ 45.5; 45.5 < T2 ≤ 80.3; and T3 > 80.3. Association between tertiles of LAP index and atherogenic lipoprotein profile were tested in a linear trend test by raw and adjusted models: age and sex (Model A) and age, sex, smoking, use of statin, fibrate, and/or hypoglycemic drugs (Model B). In addition, comparison between groups was performed by analysis of variance (ANOVA or Kruskal-Wallis – with multiple comparisons by Tukey test) after all adjustments (Model B) with significance level at p < 0.05. Results The demographic and clinical characteristics of the 351 subjects grouped by sex are shown in Table 1. The mean age of the subjects was 49.4 years for men (range: 30–72 years) and 54.4 years for women (range: 30–74 years, p < 0.001). Women were older and reported greater use of drugs than men (83.6 versus 69.8, respectively, p = 0.001), whereas higher percentage of men were smokers (p = 0.026). More than 80% of the subjects had a prior disease at the time of screening. Hypertension was the most prevalence disease in both genders (56.9% in men and 57.1% in women), which was corroborated by the high percentage of antihypertensive drug users. This profile is in concordance with elevated frequency of hypertension in father, mother or both parents of individuals (62.9% in men and 66.2% in women). Table 2 shows results of cardiovascular risk, assessed by FRS, and biochemical and anthropometric variables stratified by sex. The FRS was similar between men (13.6 points) and women (13.5 points), indicating a moderate cardiovascular risk in both groups. Men showed higher values of WC and TG, impacting directly on elevated values of LAP in comparison with women. In contrast, women had higher values of Apo AI, HDL-C and NEFAs. Both groups showed similar profile of BMI and glucose homeostasis evaluated by glucose, insulin and HOMA-IR parameters. The influence of gender on lipid metabolism was confirmed by elevated percentage of small HDL and LDL and reduced percentage of large HDL observed in men. This profile was reinforced by the increase of LDL size in men (26.9 in men versus 27.0 in women; p = 0.001) and phenotype A in women (52.3% in men versus 70.8% in women; p = 0.001). Raw and adjusted associations between LAP and other parameters were tested by tertiles (Table 3). LAP was positively associated with TC, Apo B, NEFA, glucose, insulin, and HOMA-IR and, consequently, this association increased with FRS points. Surprisingly, LAP was not corelated with LDL-C. After multiple adjustments for potential confounders (A and B models), the associations between LAP and biochemical parameters were maintained, except for Apo AI. Also, central lipid accumulation was positively associated with the percentage of intermediate and small HDL subfractions in both total (Figure 1A) and sex-stratified sample (Figures 1B, 1C) after adjustment for age, smoking, and use of statin, fibrate and hypoglycemic drugs. Similar results were found for small LDL, i.e., individuals in lowest, in the middle and in the highest tertile showed about 1.5%, 2.3% and 7.5% of small LDL, respectively (p < 0.001) (Figure 2Aii). Higher differences were seen in men (Figure 1Bi). LDL size and percentage of large HDL were both negatively associated with LAP. In total sample, this difference was nearly 10 points for large HDL – 34.2% in T1 and 24.5% in T3 (Figures 1Ai, Bi, Ci). Associations between LAP index and large LDL were found in men (Figure 2Bi), but not in total sample nor in women, demonstrating a sex-dependent relationship for this subfraction. Discussion Based on this cross-sectional study, LAP has a significant association with classical and new cardiovascular biomarkers. These associations were especially important when LAP index was corelated to size of the LDL and HDL particles. Previously, Kahn and Valdez 8 evaluated a cross-sectional sample from the NHANES III and reported that individuals with high WC and TG levels were more likely to show inadequate levels of HDL-C, Apo B, fasting insulin, and glucose. Later, Kahn 11 confirmed that the LAP was superior to BMI in indicating adults with diabetes mellitus and for predicting imbalance in glucometabolic variables (HOMA‑IR, fasting glucose, and glycated hemoglobin). Similar results were found in studies conducted in other countries, in 341