- Email: [email protected]
Comparability of two different polyacrylamide gel electrophoresis methods for the classification of LDL pattern type
Clinica Chimica Acta 413 (2012) 251–257
Contents lists available at SciVerse ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Comparability of two different polyacrylamide gel electrophoresis methods for the classification of LDL pattern type Celia Bañuls a, b, Lorena Bellod a, b, Ana Jover a, b, Maria Luisa Martínez-Triguero c, Víctor Manuel Víctor a, b, d, e, Milagros Rocha a, b, e, Antonio Hernández-Mijares a, b, f,⁎ a
Service of Endocrinology, University Hospital Dr. Peset, Valencia, Spain Dr. Peset Hospital Research Foundation, Valencia, Spain Unit of Clinical Analysis, University Hospital La Fe, Valencia, Spain d Department of Physiology, Faculty of Medicine, University of Valencia, Spain e CIBER CB06/04/0071 research group, CIBER Hepatic and Digestive Diseases, University of Valencia, Spain f Department of Medicine, Faculty of Medicine, University of Valencia, Spain b c
a r t i c l e
i n f o
Article history: Received 23 June 2011 Received in revised form 29 September 2011 Accepted 30 September 2011 Available online 6 October 2011 Keywords: Small, dense LDL Cardiovascular risk Polyacrylamide gradient gel electrophoresis Polyacrylamide gradient tube electrophoresis Triglycerides
a b s t r a c t Background: The measurement of small dense low-density lipoprotein (sdLDL) particles is relevant when assessing cardiovascular risk. However, there is as yet no referenced method for the determination of LDL subfractions or a standardized comparison of the methods currently available. Therefore, the aim of this study was to compare the pattern of LDL particles measured by polyacrylamide tube gel electrophoresis (PTGE) and polyacrylamide gradient gel electrophoresis (PGGE) and to correlate the results with triglyceride concentration. Materials and methods: Serum samples were collected from 177 patients. Lipid profile and LDL particle size were assessed using PTGE and PGGE. Results: Pearson correlation and kappa index revealed a very good agreement between the methods. There was 81.3% concordance for classification of sdLDL particles and 97.2% concordance for classification of large LDL when PTGE and PGGE were compared. LDL size correlated with triglyceride in subjects with triglyceride levels N116 mg/dl, pointing to a high CAD risk, as reflected by their higher prevalence of pattern B. Conclusions: PTGE correlates favourably and is in very good agreement with PGGE. The determination of LDL particle size may be an appropriate analytical procedure to estimate CAD risk in patients with high triglyceride levels. © 2011 Elsevier B.V. All rights reserved.
1. Introduction The risk of atherosclerosis has been related with abnormalities in plasma lipoprotein concentrations. Specifically, a connection has been demonstrated between high plasma levels of low-density lipoprotein (LDL) cholesterol and an increased risk of coronary heart disease (CHD) [1]. Thus, it is logical that LDL cholesterol is the primary therapeutic target for coronary artery disease (CAD) prevention. Nevertheless, an increasing amount of research over the past decade has been devoted to the heterogeneity of LDL particles and the atherogenicity of lipids and lipoproteins other than the LDL type. It is well known that LDL particles are a heterogeneous population that varies with respect to lipoprotein composition, size and metabolism. LDL particle size has been the focus of attention because of their atherogenic potential. Typically, individuals with large LDL particles are classified as pattern A, whereas individuals with small and dense LDL (sdLDL) particles are classified as pattern B and ⁎ Corresponding author at: Service of Endocrinology, University Hospital Dr. Peset, Av. Gaspar Aguilar 90, 46017 Valencia, Spain. Tel.: +34 961622492; Fax: +34 961622492. E-mail address: [email protected] (A. Hernández-Mijares). 0009-8981/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2011.09.047
are at a greater risk of developing CAD [2,3]. Therefore, the predominance of sdLDL particles is considered an emerging risk factor for CAD by the National Cholesterol Education Program Adult Treatment Panel III [4]. These particles are more easily concentrated in arterial walls, are more prone to oxidation, and have a reduced affinity for LDL receptors when compared with larger LDL particles [5]. When estimating cardiovascular disease (CVD) risk measurement of LDL subfractions may provide additional predictive power to LDL cholesterol concentration measurement, alone or in combination with assessment of other risk factors. For instance, individuals with smaller LDL particle size will be at higher risk of CAD than those with a similar normal/borderline elevated LDL cholesterol concentration but larger LDL particle size. Therefore, in these cases, the risk of atherosclerosis is better indicated by LDL particle size [6]. Despite the established role of sdLDL particles in the development of atherosclerosis, there is no reference method for determining LDL subfractions, since currently available analytical approached require expensive instrumentation and are timeand labour-intensive, and thus are not easily applicable in routine clinical practice [7]. Currently, methods for separating LDL subfractions include density gradient ultracentrifugation [8], nuclear magnetic resonance
252
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
[9], and non-denaturing polyacrylamide gradient gel electrophoresis (PGGE) [10]. Unfortunately, there is no standardization among the different methods employed, and comparability in terms of results is not extrapolated. In fact, the results of a systematic review have raised controversy, finding a wide range of variability of concordance (from 7 to 94%) in the classification of LDL particle patterns [11]. An alternative and less demanding approach to the study of LDL subclasses is a modified PGGE technique called polyacrylamide tube gel electrophoresis (PTGE), which has been commercialised by the Lipoprint® System (Quantimetrix, Redondo Beach, CA, USA) [12]. It measures particle size and allows sdLDL fractions to be quantified, and is relatively simple to operate and fast when compared with the PGGE. It is approved by the Food and Drug Administration (FDA) and uses a standardized method to separate LDL particles into 7 subfractions on the basis of size. As the gels are prepared by the manufacturer, it is technically simpler, less resourceintensive, and more conducive to routine laboratory testing than traditional gel electrophoresis [13]. Only a few studies have compared PTGE and PGGE [12–14], and concordance rates vary according to LDL patterns. Therefore, the aim of this study was to compare the pattern of LDL particles measured by PTGE on the one hand and PGGE on the other and to associate the results of the former method with lipid parameters of CVD risk that are known to be affected by LDL particle size. 2. Materials and methods 2.1. Subjects Patients were recruited in the Service of Endocrinology and Nutrition of University Hospital Dr Peset (Valencia, Spain). Blood samples from 177 patients were collected after 12 h overnight fasting. Patients were excluded if they presented a neoplastic, renal or liver disease or were suffering from a recent trauma or major surgery. Pregnant and lactating women were also excluded. In order to separate serum from blood cells, venous blood samples were centrifuged at 2000 g for 15 min at 4 °C. Freshly separated serum was employed to determine lipid profile, and the remaining aliquots of serum were stored at −80 °C until determination of LDL subfractions by the two aforementioned methods in our Lipid Research Unit. The study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by our hospital's Ethics Committee. Written informed consent was obtained from all patients. 2.2. Lipid parameters Total cholesterol and triglycerides were measured by means of enzymatic assays [15,16], and high-density lipoprotein (HDL) cholesterol concentrations were recorded with a Beckman LX-20 autoanalyzer (Beckman Coulter, La Brea, CA, USA) using a direct method [17]. The intraserial variation coefficient was b3.5% for all determinations. LDL cholesterol concentration was calculated using the method of Friedewald [18]. Atherogenic index of plasma (AIP) was obtained by calculating the logarithm of the ratio of plasma concentration of triglycerides to HDL cholesterol. Apolipoprotein B (Apo B) was determined by immunonephelometry (Dade Behring BNII, Marburg, Germany) with an intraassay variation coefficient of b5.5%. 2.3. Electrophoretic methods 2.3.1. Polyacrylamide gradient gel electrophoresis LDL particle size was determined by polyacrylamide gradient gel electrophoresis (2–16%), which was freshly cast in the laboratory according to the method described by Nichols et al., with slight modifications [19]. Serum samples (10 μl), stained with a solution of ethylene-glycol and 0.1% (w/v) of Sudan black (Sigma-Aldrich Corporation St. Louis, MO
63178, USA), were applied to the gel in a final concentration of 10% sucrose. Electrophoresis was performed in a refrigerated cell (4–8 °C) for a prerun of 60 min at 120 V, followed by 30 min at 20 V, 30 min at 70 V, and 15 h at 90 V. Two standard serums containing 3 LDL fractions with diameters of 28.35 nm, 26.15 nm and 24.44 nm were used as controls. The gels were scanned, and migration distances (from the top of the gel to the most prominent band) were measured. The predominant LDL particle diameter of each sample was calculated from a calibration line. LDL particle subclasses were classified as predominantly small LDL or pattern B (diameter b25.5 nm) and non-small LDL (pattern A, diameter ≥25.5 nm) [2]. Both intra- and inter-gel variation coefficients were below 1%. Representative appearance of LDL bands is shown in Fig. 1A. 2.3.2. Polyacrylamide tube gel electrophoresis LDL subfractions were separated using the Quantimetrix Lipoprint® system (Quantimetrix Corporation, Redondo Beach, CA, USA) [12], according to the procedure set out by the manufacturer. In brief, 25 μl of sample were mixed by inversion with 200 μl of Lipoprint loading gel. This mix was placed on the upper part of precast high resolution 3% polyacrylamide gel tubes. After 30 min of photopolymerization at room temperature in front of a fluorescent light source, samples were electrophoresed for 60 min at a constant current of 3 mA for each gel tube until the HDL subfraction had migrated a distance of approximately 1 cm from the bottom of the tube. Electrophoresis was followed by placing the tubes in the dark for 30 min before performing densitometry at 610 nm. According to electrophoretic mobility based on retardation factor (Rf), very-low density lipoprotein (VLDL) remains in the origin (Rf = 0.0), whereas HDL migrates to the front (Rf = 1.0). In between, up to 7 LDL bands can be detected. The LDL1 and LDL2 bands correspond with large, buoyant LDL particles, whereas bands LDL3 to LDL7 correspond with small dense LDL particles. The Rf of LDL subfraction is equal to the distance between VLDL and LDL subfraction bands divided by the distance between VLDL and HDL bands. According to this method, LDL subclasses designated as small have an Rf of more than 0.40, those designated as intermediate have an Rf of 0.38 to 0.40, and those designated as large have an Rf of less than 0.38. Therefore, according to the LDL electrophoretic profile, 2 patterns can be defined: pattern A, with normal total cholesterol mass of the sdLDL subfractions; and pattern B, in which total cholesterol mass of the sdLDL subfractions is intermediate-low. This system does not measure LDL particle size, but given that older systems do, the Lipoprint System provides an estimate by means of the algorithm developed by Kazumi [20]. The Liposure® (Quantimetrix Corporation) was used as quality control. Raw data obtained from the densitometer were imported and analysed using a computerized method developed for the Quantimetrix Lipoprint system (Lipoware software). Subfractions were identified and quantified using NIH image program version 1.62 (Bethesda, MD, USA) for research use. The program partitions bands into discrete segments, and the relative area under the curve is calculated for each lipoprotein band. It also calculates the cholesterol concentration for each lipoprotein fraction according to a total cholesterol value obtained for each sample using the aforementioned method. The average particle size reported by the Lipoprint profile is the weighted average (calculated from the area under the curve for each subfraction) of the particle size of all LDL peaks present in the sample. Based on this, the size cut-off works out more than 26.8 nm for pattern A (normal LDL size) and to be equal to or less than this threshold for pattern B. Typical migration pattern of large buoyant LDL and small dense LDL are represented in Fig. 1B and C, respectively. 2.4. Statistical analysis Statistical analyses were performed using SPSS 15.0 software (SPSS Statistics Inc., Chicago, IL, USA). Continuous variables were expressed as
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
253
A
B
VLDL IDL LDL subfractions
HDL
C
VLDL IDL LDL subfractions
HDL
Fig. 1. Representative image of LDL bands in PGGE (A) and typical migration pattern of large buoyant LDL (B) and small dense LDL in PTGE (C), with the corresponding scanned and computerized lipoprotein profile. A. Lanes 1, 2, 4, 7 and 10 show samples with pattern A, while lanes 3, 8 and 9 show samples with pattern B. Lanes 5 and 6 are two standard serums containing 3 LDL fractions with diameters of 28.35 nm, 26.15 nm (C1) and 24.44 nm (C2).
mean and standard deviation or as median and 25th and 75th percentiles for parametric and non-parametric data, respectively. Qualitative data were expressed as percentages. Comparison of the groups was performed using a one-way ANOVA and Chi-squared test for qualitative variables. Correlation between variables was determined using Spearman's correlation coefficient. Concordance between the different methods used to measure LDL subfractions was assessed using the kappa index (K). Values of 0.21–0.40, 0.41–0.60, 0.61–0.80 and 0.81–1.0 showed fair, moderate, good and very good
concordance, respectively [21]. The individual differences between the two methods were calculated and plotted, as described by Bland and Altman [22]. This method assesses the presence of bias when the difference between two methods is not equal across increasing mean levels of the parameter under study. To further clarify the differences between LDL size revealed by PTGE and PGGE, we modelled the relation between the difference and the mean of both measurements using linear regression. All the tests used a confidence interval of 95%, and differences were considered significant when p b 0.05.
254
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
3. Results The characteristics of the study population are shown in Table 1. A total of 177 subjects (27.1% males, 72.9% females) were selected in order to provide a wide range of triglyceride and lipid levels, ensuring a distribution of LDL subfractions. Comparison of PTGE and PGGE with respect to LDL particle size and classification of different LDL patterns is shown in Fig. 2 and Table 2. In relation to LDL particle size, both methods were positively correlated, providing a correlation coefficient of 0.568, which proved to be highly significant (pb 0.001) (Fig. 2A). To detect the between-method bias for measuring LDL size, the absolute difference was plotted against the mean for each pair of measurements (Fig. 2B). The mean difference between PTGE and PGGE was 1.2 nm (with PGGE being smaller). The 95% limits of agreement at this point were 0.06 and 2.34, indicating that the differences between the two methods can be expected to fall within this range. The Bland-Altman plot indicates that the absolute difference between the two methods increases with increasing average LDL size. The linear regression line shown in the figure deviated significantly from a horizontal line (pb 0.001). In addition to this, classification of LDL pattern was based on electrophoretic mobility for both methods. For PTGE, pattern B was defined when diameter was ≤ 26.8 nm and pattern A was confirmed when diameter was N26.8 nm. For PGGE, pattern B and pattern A was defined when diameter was b 25.5 nm and ≥25.5 nm, respectively. PTGE showed an agreement of 81.3% concordance for classification of pattern B when compared with PGGE. In the case of pattern A, PTGE had an agreement of 97.2% concordance when compared with PGGE. Furthermore, a very good agreement was detected between PTGE and PGGE after evaluating the data using the kappa statistic (weight K =0.81; very good concordance between 0.81 and 1.00 [21]) (Table 2). Triglyceride and ApoB concentrations and AIP correlated with LDL particle size when measured by PTGE and PGGE methods (Table 3). Good correlation was confirmed for triglycerides (r= −0.705, p b 0.001 and r = −0.484, p b 0.001 for PTGE and PGGE, respectively), Apo B (r= −0.532, p b 0.001 and r = −0.244, p = 0.001 for PTGE and PGGE, respectively) and AIP (r= −0.607, p b 0.001 and r = −0.435, p b 0.001 for PTGE and PGGE, respectively), although the highest correlation coefficients for all parameters were associated with PTGE. For a deeper analysis of correlations, we stratified our population using triglyceride tertiles. The results showed that, for triglycerides and AIP, correlation was only significant in subjects of the third tertile; that is, those Table 1 Baseline characteristics and lipid profile of study population.
Sample size (n) Male/Female (n) Normolipidemia (%) Cases of hyperlipidemia (%) TC N 200 mg/dl Triglycerides N 200 mg/dl TC and triglycerides N 200 mg/dl Lipid-lowering therapy (%) Cases of metabolic syndrome (%) Cases of type 2 diabetes (%) Post-menopausal women (%) Age (years) BMI (Kg/m2) TC (mg/dl) LDLc (mg/dl) HDLc (mg/dl) Triglycerides (mg/dl) Apo B (mg/dl) AIP
Mean ± SD
Range
177 48/129 51
————— —————
47 11 8.5 10 26 21 39 43 ± 19 27.4 ± 6.1 196 ± 41 126 ± 35 48 ± 15 93 (63,146) 92 ± 26 0.319 ± 0.337
————— ————— ————— ————— ————— 18–92 17.6–48.1 105–341 47–245 21–93 28–591 46–178 − 0.320–1.450
Data are expressed as mean ± standard deviation for parametric data or as median (25th and 75th percentiles) for non-parametric data. Abbreviations: BMI: body mass index; TC: total cholesterol, LDLc: low-density lipoprotein cholesterol, HDLc: high-density lipoprotein cholesterol; Apo B: Apolipoprotein B, AIP: atherogenic index of plasma = (log (triglycerides/HDL cholesterol)).
with levels higher than 116 mg/dl. Furthermore, these subjects showed a greater risk of CVD, as reflected by their higher levels of total cholesterol and LDL cholesterol and lower levels of HDL) than subjects in the first and second tertiles (data not shown). Finally, we analysed diameter and pattern of LDL –both measured by the PTGE and PGGE methods — according to triglyceride tertiles. We found a significant reduction in the diameter of LDL particles in patients in the third tertile when compared with those in the first and second tertiles (Fig. 3A). In line with this, pattern A was predominant among subjects in the first and second tertiles, with only six for PTGE and five for PGGE showing pattern B (0% and 9.7% in the first and second tertiles, respectively, with PTGE and 0% and 8.1% in the first and second tertiles, respectively, with PGGE), while pattern B was predominant among subjects in the third tertile (highest triglyceride levels) (41.4% with PTGE and 46.6% with PGGE) (Fig. 3B). 4. Discussion In the last decade, there has been increasing interest in characterizing and measuring LDL subfractions due to their involvement in the atherosclerotic process. Among the different approaches currently employed, PGGE, preceded by ultracentrifugation and nuclear magnetic resonance spectroscopy, is a well-established and precise method for the separation of LDL subfractions, though it is time-consuming and technically demanding. As new methods are introduced (for example, PTGE), standardization becomes increasingly important, as the principles on which each approach is based can vary, and there is little information to compare the various analytical techniques available. As we have mentioned in Material & Methods, the PTGE method developed by Quantimetrix Lipoprint® system (Quantimetrix Corporation, Redondo Beach, CA, USA) allows the separation of LDL subfractions into 7 bands based primarily on LDL electrophoretic mobility. LDL subfractions 1 and 2 have been designated as large LDL, and subfractions 3 to 7 as sdLDL, with a subsequent reduction in LDL size as fraction number increases. We analyzed LDL particle size in 177 samples using two different analytical methods: PTGE and PGGE. Both methods were highly correlated, though it must be pointed out that correlation is a measure of the strength of a relation (or association) rather than agreement between two factors. Due to the underlying statistical assumption of correlation coefficients, interpretation of test accuracy using these values can be misleading, as a correlation coefficient depends on the range of values in the sample, and the correlation is likely to be stronger when the range is wide rather than narrow. Furthermore, it is difficult to assign a correlation coefficient threshold value for widespread acceptance of a test. In this context, we focused on a diagnostic classification of patient LDL pattern type that goes beyond the quantitative data of particle size, as the latter parameter is often not comparable due to differences between the methodologies employed. Our results indicate that 94.4% of the study population had the same LDL particle pattern, with only 10 patients being misclassified. Despite this, there was very good agreement between the 2 methods, with a weighted kappa index of 0.81. Of the 32 samples classified by PGGE as sdLDL particles, 6 were misclassified by PTGE as large LDL particles (81.3% concordance), and of the 145 samples classified by PGGE as large, 4 were misclassified by PTGE as small (97.2% concordance). Consequently, this method of analysis appears to permit a simplified categorization of patients as displaying pattern A or pattern B, with good correlation and concordance with PGGE. To our knowledge, only three studies have addressed this topic previously, and all three classified LDL subfractions as small, intermediate or large based on different cut-off points. In accordance with our results, Hirany et al. [13] reported good agreement between PTGE and PGGE methods after evaluating the data using kappa statistics (weighed K =0.78), with the former method showing an agreement of 92%, 33% and 77% concordance for classification of small, intermediate and large LDL subfractions, respectively, when compared with PGGE. Similarly, another study reported 100%, 64%, and 95% agreement for classification of
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
A
255
B 28
LDL size diference (nm)
LDL diameter by PGGE (nm)
4 29 r=0.568, p<0.001
27 26 25 24 25
26
27
28
LDL diameter by PTGE (nm)
r = -0.617, p<0.001
3 2 1 0 -1 -2 24.5
25.5
26.5
27.5
Average LDL size (nm)
Fig. 2. LDL size analyzed by PGGE plotted against LDL size measured by PTGE (A) and Bland-Altman plot of absolute differences between LDL size determined by PTGE and PGGE vs the average of the two methods, with the mean of the differences (solid line) and mean ± 2SD limits (broken line) (B). Abbreviations: PTGE: polyacrylamide tube gel electrophoresis; PGGE: polyacrylamide gradient gel electrophoresis.
small, intermediate and large LDL subfractions, respectively, although the kappa index was not provided [12]. In a more recent systematic review, the abovementioned concordance percentages were re-examined, showing a noticeable change from 100% to 88% and from 95% to 24% for small and large LDL subfractions, respectively [11]. These differences were based on the assignment of subjects with an intermediate pattern. Finally, Ensign and coworkers [14] compared LDL pattern as determined by different technologies – PGGE, ultracentrifugation, nuclear magnetic resonance and PTGE – and demonstrated disparities among the methods, showing that complete agreement with respect to LDL subclass phenotype occurred in only 8% of the samples analyzed, reporting a mere 41% agreement between PTGE and PGGE with respect to the classification of LDL pattern (38% for large LDL and 3% for sdLDL). The small sample sizes of these studies [12,14] and/or their categorization of LDL in three patterns (the intermediate pattern being most prone to misclassification) could justify the variability in concordance indexes reported. Additionally, computer software for automated analysis of the bands was not available at the time the studies were published; therefore, each system would have had a different cut-off point for the classification of each pattern. Currently, the automated system of PTGE supplied by Quantimetrix (Lipoware software) allows a standardized analysis of LDL subfraction. In our Lipid Research Unit, we have observed that the PTGE method is precise and compares favourably with the PGGE method when employed for diagnostic classification of LDL particle type. Thus, we would affirm that both analytical methods can be used interchangeably. Additional advantages of PTGE with respect to PGGE is that it is faster and easier to perform when assessing large numbers of samples, and therefore could be of great practicality in clinical practice. However, PTGE is a relatively expensive procedure and may not be suitable for all patients. As a consequence, its use should be limited to patients in whom the determination of LDL particle size is likely to provide a real assessment of CAD risk. Table 2 Comparability of polyacrylamide tube gel electrophoresis and polyacrylamide gradient gel electrophoresis as methods of LDL subfraction determination. PGGE
Pattern A (LDL size ≥25.5 nm) Pattern B (LDL size b 25.5 nm)
Number of cases
PTGE Pattern A
Pattern B
(LDL size ≥26.8 nm)
(LDL size b26.8 nm)
Concordance (%)*
145
141
4
97.2
32
6
26
81.3
*weighed kappa of 0.81. Abbreviations: PTGE: polyacrylamide tube gel electrophoresis; PGGE: polyacrylamide gradient gel electrophoresis.
An increased proportion of sdLDL particles has been associated with marked alterations in serum lipoprotein and lipid levels, in particular elevated triglyceride concentrations. In accordance with that reported by other studies [12,13,20,23,24], we found that triglyceride and Apo B concentration –a indicator of the total number of LDL particles [25] – and AIP – considered a surrogate of sdLDL particle size [26]– were closely associated with LDL particle size. By dividing triglyceride levels into tertiles (firstb 68 mg/dl, second 68–116 mg/dl and thirdN 116 mg/dl), we were able to observe that the correlation of triglyceride and AIP with LDL particle size disappeared in the lower (first and second) tertiles, showing a good correlation, and one which was even better than in the general population, when only subjects in the third tertile were considered. Thus, triglyceride levels did not predict the LDL particle size of subjects in the first and second tertiles. Other lipoproteins may have been responsible for the variation of LDL particle size in our population, although this parameter was relatively constant, since only six of the 119 subjects was classified as pattern B (5.0%). Thus, Apo B showed a good correlation with LDL size (measured by PTGE) in all tertiles, which is in accordance with the results of a recent study that compared the PTGE classification method with measurement of Apo B and direct sdLDL [27]. In contrast, we detected a pattern that was predominantly of the B type (41.4%) and a more atherogenic lipid profile among patients in the third tertile that was associated with high total and LDL cholesterol (borderline high-risk), Apo B levels and AIP and low HDL cholesterol levels. In this way, our results suggest that determining LDL subfractions in patients with higher-than-normal triglyceride levels
Table 3 Correlation of triglycerides, atherogenic index and Apo B of plasma with LDL particle size measured by polyacrylamide gel gradient electrophoresis and polyacrylamide tube gradient electrophoresis. PTGE
Total triglycerides (mg/dl) 1st tertile 2nd tertile 3rd tertile AIP (log (triglycerides/HDLc) 1st tertile 2nd tertile 3rd tertile Apo B (mg/dl) 1st tertile 2nd tertile 3rd tertile
PGGE
r*
p
r*
p
− 0.705 0.086 − 0.123 − 0.676 − 0.607 0.055 − 0.123 − 0.636 − 0.532 − 0.305 − 0.565 − 0.361
b 0.001 0.522 0.339 b 0.001 b 0.001 0.686 0.342 b 0.001 b 0.001 0.021 b 0.001 0.005
− 0.484 0.129 − 0.028 − 0.527 − 0.435 0.094 0.088 − 0.552 − 0.244 0.051 − 0.321 0.060
b 0.001 0.338 0.830 b 0.001 b 0.001 0.487 0.496 b 0.001 0.001 0.705 0.012 0.652
*Pearson correlation. 1st tertile included 57 subjects with triglyceride levels b68 mg/dl, 2nd tertile included 62 subjects with triglyceride levels between 68 mg/dl and 116 mg/dl and 3 rd tertile included 58 subjects with triglyceride levels N 116 mg/dl.
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
A LDL diameter (nm)
B
PTGE
27.5 a
a
27.3 27.1
b
26.9 26.7 26.5
<68
68-116
PGGE
26.5
LDL diameter (nm)
256
26.3
a
25.9 25.7 25.5
>116
a
26.1
b <68
Pattern A
Pattern B
Pattern B
58.6 100
90.3 414 41.4
0 <68
9.7 68-116
>116
Pattern A
100
50
68-116
Triglycerides (mg/dl)
>116
% of cases by PGGE
% of cases by PTGE
Triglycerides (mg/dl)
100
53.4 50
100
91.1
<68
8.1 68-116
66 0
Triglycerides (mg/dl)
>116
Triglycerides (mg/dl)
Fig. 3. Distribution of LDL diameter and pattern types measured by PTGE (A) and PGGE (B) according to triglyceride tertiles. Data are expressed as mean ± standard error of 57, 62 and 58 subjects for first, second and third tertile, respectively. Unlike superscript letters were significantly different when different tertiles were compared using one-way ANOVA and a Student-Newman-Keuls test as post hoc. Abbreviations: PTGE: polyacrylamide tube gel electrophoresis; PGGE: polyacrylamide gradient gel electrophoresis.
(N150 mg/dl) could be an important tool for assessing CAD risk, as this risk increases when values exceed 116 mg/dl. Considered as a whole, our results show that PTGE correlates favourably and is in good agreement with PGGE when used for the diagnostic classification of LDL pattern type. This evidence has led us to employ both analytical methods interchangeably in our Lipid Research Unit. Additionally, the determination of LDL particle size would seem to be an appropriate analytical procedure for estimating CAD risk in patients with borderline high-risk levels of total and LDL cholesterol and with triglyceride levels higher than 116 mg/dl. List of abbreviations AIP atherogenic index of plasma Apo B apolipoprotein B CAD coronary artery disease CHD coronary heart disease CVD cardiovascular disease FDA Food and Drug Administration K kappa index HDL high-density lipoprotein LDL low-density lipoprotein PGGE polyacrylamide gradient gel electrophoresis PTGE polyacrylamide tube gel electrophoresis Rf retardation factor sdLDL small dense low-density lipoprotein VLDL very-low density lipoprotein
Conflict of interest The authors claim no conflict of interest associated with the manuscript.
Acknowledgements Sources of financial support: The study has been supported by grants PS09/01025 and PI10/1195 from Fondo de Investigación Sanitaria (FIS) and AP-192/11 from Regional Ministry of Health of Valencian Community. M Rocha is a recipient of Miguel Servet contract from FIS (CP10/00360). VM Víctor is a recipient of Regional Ministry of Health of Valencian Community and Carlos III Health Institute contract (CES10/030). We kindly thank B Normanly and I Soria-Cuenca for their contribution to the present study.
References [1] Singh BK, Mehta JL. Management of dyslipidemia in the primary prevention of coronary heart disease. Curr Opin Cardiol 2002;17:503–11. [2] Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss RM. Lowdensity lipoprotein subclass patterns and risk of myocardial infarction. JAMA 1988;260:1917–21. [3] Krauss RM, Siri PW. Metabolic abnormalities: triglyceride and low-density lipoprotein. Endocrinol Metab Clin North Am 2004;33:405–15. [4] Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA 2001;285:48–97. [5] Rizzo M, Berneis K. Low-density lipoprotein size and cardiovascular prevention. Eur J Intern Med 2006;17:77–80. [6] Otvos JD, Mora S, Shalaurova I, Greenland P, Mackey RH, Goff Jr DC. Clinical implications of discordance between low-density lipoprotein cholesterol and particle number. J Clin Lipidol 2011;5:105–13. [7] Hirano T, Ito Y, Yoshino G. Measurement of small dense low-density lipoprotein particles. J Atheroscler Thromb 2005;12:67–72. [8] Griffin BA, Caslake MJ, Yip B, Tait GW, Packard CJ, Shepherd J. Rapid isolation of low density lipoprotein (LDL) subfractions from plasma by density gradient ultracentrifugation. Atherosclerosis 1990;83:59–67. [9] Otvos JD, Jeyarajah EJ, Bennett DW. Quantification of plasma lipoproteins by proton nuclear magnetic resonance spectroscopy. Clin Chem 1991;37:377–86.
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257 [10] Austin MA, Hokanson JE, Brunzell JD. Characterization of lowdensity lipoprotein subclasses: methodological approaches and clinical relevance. Curr Opin Lipidol 1994;5:395–403. [11] Chung M, Lichtenstein AH, Ip S, Lau J, Balk EM. Comparability of methods for LDL subfraction determination: A systematic review. Atherosclerosis 2009;205:342–8. [12] Hoefner DM, Hodel SD, O'Brien JF, et al. Development of a rapid, quantitative method for LDL subfractionation with use of the Quantimetrix Lipoprint LDL System. Clin Chem 2001;47:266–74. [13] Hirany SV, Othman Y, Kutscher P, Rainwater DL, Jialal I, Devaraj S. Comparison of low-density lipoprotein size by polyacrylamide tube gel electrophoresis and polyacrylamide gradient gel electrophoresis. Am J Clin Pathol 2003;119:439–45. [14] Ensign W, Hill N, Heward CB. Disparate LDL phenotypic classification among 4 different methods assessing LDL particle characteristics. Clin Chem 2006;52:1722–7. [15] Allain CC, Poon LS, Chan CS, Richmond W, Fu PC. Enzymatic determination of total serum cholesterol. Clin Chem 1974;20:470–5. [16] Bucolo G, David H. Quantitative determination of serum triglycerides by the use of enzymes. Clin Chem 1973;19:476–82. [17] Sugiuchi H, Uji Y, Okabe H, et al. Direct measurement of high-density lipoprotein cholesterol in serum with polyethylene glycol-modified enzymed and sulfates alphacyclodextrin. Clin Chem 1995;41:717–23. [18] Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of preparative ultracentrifuge. Clin Chem 1972;18:499–502. [19] Nichols AV, Krauss RM, Musliner TA. Nondenaturing polyacrylamide gradient gel electrophoresis. Methods Enzymol 1986;128:417–31.
257
[20] Kazumi T, Kawaguchi A, Hozumi T, et al. Low density lipoprotein particle diameter in young, nonobese, normolipidemic Japanese men. Atherosclerosis 1999;142: 113–9. [21] Altman DG. Some common problems in medical research. In: Altman DG, editor. Practical statistics for medical research. New York: Chapman and Hall; 1991. p. 396–437. [22] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [23] Lamarche B, Tchernof A, Moorjani S, et al. Small, dense low-density lipoprotein particles as a predictor of the risk of ischemic heart disease in men. Prospective results from the Québec Cardiovascular Study. Circulation 1997;95:69–75. [24] Gardner CD, Fortmann SP, Krauss RM. Association of small low-density lipoprotein particles with the incidence of coronary artery disease in men and women. JAMA 1996;276:875–81. [25] Contois JH, McConnell JP, Sethi AA, et al. Apolipoprotein B and cardiovascular disease risk: Position statement from the AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Clin Chem 2009;55:407–19. [26] Dobiásová M, Frohlich J. The plasma parameter log (TG/HDL-C) as an atherogenic index: correlation with lipoprotein particle size and esterification rate in apoBlipoprotein-depleted plasma (FER(HDL)). Clin Biochem 2001;34:583–8. [27] Vandermeersch A, Ameye S, Puype D, Petitjean D, De Buyzere M, Langlois MR. Estimation of the low-density lipoprotein (LDL) subclass phenotype using a direct, automated assay of small dense LDL-cholesterol without sample pretreatment. Clin Chim Acta 2010;411:1361–6.
Contents lists available at SciVerse ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Comparability of two different polyacrylamide gel electrophoresis methods for the classification of LDL pattern type Celia Bañuls a, b, Lorena Bellod a, b, Ana Jover a, b, Maria Luisa Martínez-Triguero c, Víctor Manuel Víctor a, b, d, e, Milagros Rocha a, b, e, Antonio Hernández-Mijares a, b, f,⁎ a
Service of Endocrinology, University Hospital Dr. Peset, Valencia, Spain Dr. Peset Hospital Research Foundation, Valencia, Spain Unit of Clinical Analysis, University Hospital La Fe, Valencia, Spain d Department of Physiology, Faculty of Medicine, University of Valencia, Spain e CIBER CB06/04/0071 research group, CIBER Hepatic and Digestive Diseases, University of Valencia, Spain f Department of Medicine, Faculty of Medicine, University of Valencia, Spain b c
a r t i c l e
i n f o
Article history: Received 23 June 2011 Received in revised form 29 September 2011 Accepted 30 September 2011 Available online 6 October 2011 Keywords: Small, dense LDL Cardiovascular risk Polyacrylamide gradient gel electrophoresis Polyacrylamide gradient tube electrophoresis Triglycerides
a b s t r a c t Background: The measurement of small dense low-density lipoprotein (sdLDL) particles is relevant when assessing cardiovascular risk. However, there is as yet no referenced method for the determination of LDL subfractions or a standardized comparison of the methods currently available. Therefore, the aim of this study was to compare the pattern of LDL particles measured by polyacrylamide tube gel electrophoresis (PTGE) and polyacrylamide gradient gel electrophoresis (PGGE) and to correlate the results with triglyceride concentration. Materials and methods: Serum samples were collected from 177 patients. Lipid profile and LDL particle size were assessed using PTGE and PGGE. Results: Pearson correlation and kappa index revealed a very good agreement between the methods. There was 81.3% concordance for classification of sdLDL particles and 97.2% concordance for classification of large LDL when PTGE and PGGE were compared. LDL size correlated with triglyceride in subjects with triglyceride levels N116 mg/dl, pointing to a high CAD risk, as reflected by their higher prevalence of pattern B. Conclusions: PTGE correlates favourably and is in very good agreement with PGGE. The determination of LDL particle size may be an appropriate analytical procedure to estimate CAD risk in patients with high triglyceride levels. © 2011 Elsevier B.V. All rights reserved.
1. Introduction The risk of atherosclerosis has been related with abnormalities in plasma lipoprotein concentrations. Specifically, a connection has been demonstrated between high plasma levels of low-density lipoprotein (LDL) cholesterol and an increased risk of coronary heart disease (CHD) [1]. Thus, it is logical that LDL cholesterol is the primary therapeutic target for coronary artery disease (CAD) prevention. Nevertheless, an increasing amount of research over the past decade has been devoted to the heterogeneity of LDL particles and the atherogenicity of lipids and lipoproteins other than the LDL type. It is well known that LDL particles are a heterogeneous population that varies with respect to lipoprotein composition, size and metabolism. LDL particle size has been the focus of attention because of their atherogenic potential. Typically, individuals with large LDL particles are classified as pattern A, whereas individuals with small and dense LDL (sdLDL) particles are classified as pattern B and ⁎ Corresponding author at: Service of Endocrinology, University Hospital Dr. Peset, Av. Gaspar Aguilar 90, 46017 Valencia, Spain. Tel.: +34 961622492; Fax: +34 961622492. E-mail address: [email protected] (A. Hernández-Mijares). 0009-8981/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.cca.2011.09.047
are at a greater risk of developing CAD [2,3]. Therefore, the predominance of sdLDL particles is considered an emerging risk factor for CAD by the National Cholesterol Education Program Adult Treatment Panel III [4]. These particles are more easily concentrated in arterial walls, are more prone to oxidation, and have a reduced affinity for LDL receptors when compared with larger LDL particles [5]. When estimating cardiovascular disease (CVD) risk measurement of LDL subfractions may provide additional predictive power to LDL cholesterol concentration measurement, alone or in combination with assessment of other risk factors. For instance, individuals with smaller LDL particle size will be at higher risk of CAD than those with a similar normal/borderline elevated LDL cholesterol concentration but larger LDL particle size. Therefore, in these cases, the risk of atherosclerosis is better indicated by LDL particle size [6]. Despite the established role of sdLDL particles in the development of atherosclerosis, there is no reference method for determining LDL subfractions, since currently available analytical approached require expensive instrumentation and are timeand labour-intensive, and thus are not easily applicable in routine clinical practice [7]. Currently, methods for separating LDL subfractions include density gradient ultracentrifugation [8], nuclear magnetic resonance
252
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
[9], and non-denaturing polyacrylamide gradient gel electrophoresis (PGGE) [10]. Unfortunately, there is no standardization among the different methods employed, and comparability in terms of results is not extrapolated. In fact, the results of a systematic review have raised controversy, finding a wide range of variability of concordance (from 7 to 94%) in the classification of LDL particle patterns [11]. An alternative and less demanding approach to the study of LDL subclasses is a modified PGGE technique called polyacrylamide tube gel electrophoresis (PTGE), which has been commercialised by the Lipoprint® System (Quantimetrix, Redondo Beach, CA, USA) [12]. It measures particle size and allows sdLDL fractions to be quantified, and is relatively simple to operate and fast when compared with the PGGE. It is approved by the Food and Drug Administration (FDA) and uses a standardized method to separate LDL particles into 7 subfractions on the basis of size. As the gels are prepared by the manufacturer, it is technically simpler, less resourceintensive, and more conducive to routine laboratory testing than traditional gel electrophoresis [13]. Only a few studies have compared PTGE and PGGE [12–14], and concordance rates vary according to LDL patterns. Therefore, the aim of this study was to compare the pattern of LDL particles measured by PTGE on the one hand and PGGE on the other and to associate the results of the former method with lipid parameters of CVD risk that are known to be affected by LDL particle size. 2. Materials and methods 2.1. Subjects Patients were recruited in the Service of Endocrinology and Nutrition of University Hospital Dr Peset (Valencia, Spain). Blood samples from 177 patients were collected after 12 h overnight fasting. Patients were excluded if they presented a neoplastic, renal or liver disease or were suffering from a recent trauma or major surgery. Pregnant and lactating women were also excluded. In order to separate serum from blood cells, venous blood samples were centrifuged at 2000 g for 15 min at 4 °C. Freshly separated serum was employed to determine lipid profile, and the remaining aliquots of serum were stored at −80 °C until determination of LDL subfractions by the two aforementioned methods in our Lipid Research Unit. The study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all procedures involving human subjects were approved by our hospital's Ethics Committee. Written informed consent was obtained from all patients. 2.2. Lipid parameters Total cholesterol and triglycerides were measured by means of enzymatic assays [15,16], and high-density lipoprotein (HDL) cholesterol concentrations were recorded with a Beckman LX-20 autoanalyzer (Beckman Coulter, La Brea, CA, USA) using a direct method [17]. The intraserial variation coefficient was b3.5% for all determinations. LDL cholesterol concentration was calculated using the method of Friedewald [18]. Atherogenic index of plasma (AIP) was obtained by calculating the logarithm of the ratio of plasma concentration of triglycerides to HDL cholesterol. Apolipoprotein B (Apo B) was determined by immunonephelometry (Dade Behring BNII, Marburg, Germany) with an intraassay variation coefficient of b5.5%. 2.3. Electrophoretic methods 2.3.1. Polyacrylamide gradient gel electrophoresis LDL particle size was determined by polyacrylamide gradient gel electrophoresis (2–16%), which was freshly cast in the laboratory according to the method described by Nichols et al., with slight modifications [19]. Serum samples (10 μl), stained with a solution of ethylene-glycol and 0.1% (w/v) of Sudan black (Sigma-Aldrich Corporation St. Louis, MO
63178, USA), were applied to the gel in a final concentration of 10% sucrose. Electrophoresis was performed in a refrigerated cell (4–8 °C) for a prerun of 60 min at 120 V, followed by 30 min at 20 V, 30 min at 70 V, and 15 h at 90 V. Two standard serums containing 3 LDL fractions with diameters of 28.35 nm, 26.15 nm and 24.44 nm were used as controls. The gels were scanned, and migration distances (from the top of the gel to the most prominent band) were measured. The predominant LDL particle diameter of each sample was calculated from a calibration line. LDL particle subclasses were classified as predominantly small LDL or pattern B (diameter b25.5 nm) and non-small LDL (pattern A, diameter ≥25.5 nm) [2]. Both intra- and inter-gel variation coefficients were below 1%. Representative appearance of LDL bands is shown in Fig. 1A. 2.3.2. Polyacrylamide tube gel electrophoresis LDL subfractions were separated using the Quantimetrix Lipoprint® system (Quantimetrix Corporation, Redondo Beach, CA, USA) [12], according to the procedure set out by the manufacturer. In brief, 25 μl of sample were mixed by inversion with 200 μl of Lipoprint loading gel. This mix was placed on the upper part of precast high resolution 3% polyacrylamide gel tubes. After 30 min of photopolymerization at room temperature in front of a fluorescent light source, samples were electrophoresed for 60 min at a constant current of 3 mA for each gel tube until the HDL subfraction had migrated a distance of approximately 1 cm from the bottom of the tube. Electrophoresis was followed by placing the tubes in the dark for 30 min before performing densitometry at 610 nm. According to electrophoretic mobility based on retardation factor (Rf), very-low density lipoprotein (VLDL) remains in the origin (Rf = 0.0), whereas HDL migrates to the front (Rf = 1.0). In between, up to 7 LDL bands can be detected. The LDL1 and LDL2 bands correspond with large, buoyant LDL particles, whereas bands LDL3 to LDL7 correspond with small dense LDL particles. The Rf of LDL subfraction is equal to the distance between VLDL and LDL subfraction bands divided by the distance between VLDL and HDL bands. According to this method, LDL subclasses designated as small have an Rf of more than 0.40, those designated as intermediate have an Rf of 0.38 to 0.40, and those designated as large have an Rf of less than 0.38. Therefore, according to the LDL electrophoretic profile, 2 patterns can be defined: pattern A, with normal total cholesterol mass of the sdLDL subfractions; and pattern B, in which total cholesterol mass of the sdLDL subfractions is intermediate-low. This system does not measure LDL particle size, but given that older systems do, the Lipoprint System provides an estimate by means of the algorithm developed by Kazumi [20]. The Liposure® (Quantimetrix Corporation) was used as quality control. Raw data obtained from the densitometer were imported and analysed using a computerized method developed for the Quantimetrix Lipoprint system (Lipoware software). Subfractions were identified and quantified using NIH image program version 1.62 (Bethesda, MD, USA) for research use. The program partitions bands into discrete segments, and the relative area under the curve is calculated for each lipoprotein band. It also calculates the cholesterol concentration for each lipoprotein fraction according to a total cholesterol value obtained for each sample using the aforementioned method. The average particle size reported by the Lipoprint profile is the weighted average (calculated from the area under the curve for each subfraction) of the particle size of all LDL peaks present in the sample. Based on this, the size cut-off works out more than 26.8 nm for pattern A (normal LDL size) and to be equal to or less than this threshold for pattern B. Typical migration pattern of large buoyant LDL and small dense LDL are represented in Fig. 1B and C, respectively. 2.4. Statistical analysis Statistical analyses were performed using SPSS 15.0 software (SPSS Statistics Inc., Chicago, IL, USA). Continuous variables were expressed as
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
253
A
B
VLDL IDL LDL subfractions
HDL
C
VLDL IDL LDL subfractions
HDL
Fig. 1. Representative image of LDL bands in PGGE (A) and typical migration pattern of large buoyant LDL (B) and small dense LDL in PTGE (C), with the corresponding scanned and computerized lipoprotein profile. A. Lanes 1, 2, 4, 7 and 10 show samples with pattern A, while lanes 3, 8 and 9 show samples with pattern B. Lanes 5 and 6 are two standard serums containing 3 LDL fractions with diameters of 28.35 nm, 26.15 nm (C1) and 24.44 nm (C2).
mean and standard deviation or as median and 25th and 75th percentiles for parametric and non-parametric data, respectively. Qualitative data were expressed as percentages. Comparison of the groups was performed using a one-way ANOVA and Chi-squared test for qualitative variables. Correlation between variables was determined using Spearman's correlation coefficient. Concordance between the different methods used to measure LDL subfractions was assessed using the kappa index (K). Values of 0.21–0.40, 0.41–0.60, 0.61–0.80 and 0.81–1.0 showed fair, moderate, good and very good
concordance, respectively [21]. The individual differences between the two methods were calculated and plotted, as described by Bland and Altman [22]. This method assesses the presence of bias when the difference between two methods is not equal across increasing mean levels of the parameter under study. To further clarify the differences between LDL size revealed by PTGE and PGGE, we modelled the relation between the difference and the mean of both measurements using linear regression. All the tests used a confidence interval of 95%, and differences were considered significant when p b 0.05.
254
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
3. Results The characteristics of the study population are shown in Table 1. A total of 177 subjects (27.1% males, 72.9% females) were selected in order to provide a wide range of triglyceride and lipid levels, ensuring a distribution of LDL subfractions. Comparison of PTGE and PGGE with respect to LDL particle size and classification of different LDL patterns is shown in Fig. 2 and Table 2. In relation to LDL particle size, both methods were positively correlated, providing a correlation coefficient of 0.568, which proved to be highly significant (pb 0.001) (Fig. 2A). To detect the between-method bias for measuring LDL size, the absolute difference was plotted against the mean for each pair of measurements (Fig. 2B). The mean difference between PTGE and PGGE was 1.2 nm (with PGGE being smaller). The 95% limits of agreement at this point were 0.06 and 2.34, indicating that the differences between the two methods can be expected to fall within this range. The Bland-Altman plot indicates that the absolute difference between the two methods increases with increasing average LDL size. The linear regression line shown in the figure deviated significantly from a horizontal line (pb 0.001). In addition to this, classification of LDL pattern was based on electrophoretic mobility for both methods. For PTGE, pattern B was defined when diameter was ≤ 26.8 nm and pattern A was confirmed when diameter was N26.8 nm. For PGGE, pattern B and pattern A was defined when diameter was b 25.5 nm and ≥25.5 nm, respectively. PTGE showed an agreement of 81.3% concordance for classification of pattern B when compared with PGGE. In the case of pattern A, PTGE had an agreement of 97.2% concordance when compared with PGGE. Furthermore, a very good agreement was detected between PTGE and PGGE after evaluating the data using the kappa statistic (weight K =0.81; very good concordance between 0.81 and 1.00 [21]) (Table 2). Triglyceride and ApoB concentrations and AIP correlated with LDL particle size when measured by PTGE and PGGE methods (Table 3). Good correlation was confirmed for triglycerides (r= −0.705, p b 0.001 and r = −0.484, p b 0.001 for PTGE and PGGE, respectively), Apo B (r= −0.532, p b 0.001 and r = −0.244, p = 0.001 for PTGE and PGGE, respectively) and AIP (r= −0.607, p b 0.001 and r = −0.435, p b 0.001 for PTGE and PGGE, respectively), although the highest correlation coefficients for all parameters were associated with PTGE. For a deeper analysis of correlations, we stratified our population using triglyceride tertiles. The results showed that, for triglycerides and AIP, correlation was only significant in subjects of the third tertile; that is, those Table 1 Baseline characteristics and lipid profile of study population.
Sample size (n) Male/Female (n) Normolipidemia (%) Cases of hyperlipidemia (%) TC N 200 mg/dl Triglycerides N 200 mg/dl TC and triglycerides N 200 mg/dl Lipid-lowering therapy (%) Cases of metabolic syndrome (%) Cases of type 2 diabetes (%) Post-menopausal women (%) Age (years) BMI (Kg/m2) TC (mg/dl) LDLc (mg/dl) HDLc (mg/dl) Triglycerides (mg/dl) Apo B (mg/dl) AIP
Mean ± SD
Range
177 48/129 51
————— —————
47 11 8.5 10 26 21 39 43 ± 19 27.4 ± 6.1 196 ± 41 126 ± 35 48 ± 15 93 (63,146) 92 ± 26 0.319 ± 0.337
————— ————— ————— ————— ————— 18–92 17.6–48.1 105–341 47–245 21–93 28–591 46–178 − 0.320–1.450
Data are expressed as mean ± standard deviation for parametric data or as median (25th and 75th percentiles) for non-parametric data. Abbreviations: BMI: body mass index; TC: total cholesterol, LDLc: low-density lipoprotein cholesterol, HDLc: high-density lipoprotein cholesterol; Apo B: Apolipoprotein B, AIP: atherogenic index of plasma = (log (triglycerides/HDL cholesterol)).
with levels higher than 116 mg/dl. Furthermore, these subjects showed a greater risk of CVD, as reflected by their higher levels of total cholesterol and LDL cholesterol and lower levels of HDL) than subjects in the first and second tertiles (data not shown). Finally, we analysed diameter and pattern of LDL –both measured by the PTGE and PGGE methods — according to triglyceride tertiles. We found a significant reduction in the diameter of LDL particles in patients in the third tertile when compared with those in the first and second tertiles (Fig. 3A). In line with this, pattern A was predominant among subjects in the first and second tertiles, with only six for PTGE and five for PGGE showing pattern B (0% and 9.7% in the first and second tertiles, respectively, with PTGE and 0% and 8.1% in the first and second tertiles, respectively, with PGGE), while pattern B was predominant among subjects in the third tertile (highest triglyceride levels) (41.4% with PTGE and 46.6% with PGGE) (Fig. 3B). 4. Discussion In the last decade, there has been increasing interest in characterizing and measuring LDL subfractions due to their involvement in the atherosclerotic process. Among the different approaches currently employed, PGGE, preceded by ultracentrifugation and nuclear magnetic resonance spectroscopy, is a well-established and precise method for the separation of LDL subfractions, though it is time-consuming and technically demanding. As new methods are introduced (for example, PTGE), standardization becomes increasingly important, as the principles on which each approach is based can vary, and there is little information to compare the various analytical techniques available. As we have mentioned in Material & Methods, the PTGE method developed by Quantimetrix Lipoprint® system (Quantimetrix Corporation, Redondo Beach, CA, USA) allows the separation of LDL subfractions into 7 bands based primarily on LDL electrophoretic mobility. LDL subfractions 1 and 2 have been designated as large LDL, and subfractions 3 to 7 as sdLDL, with a subsequent reduction in LDL size as fraction number increases. We analyzed LDL particle size in 177 samples using two different analytical methods: PTGE and PGGE. Both methods were highly correlated, though it must be pointed out that correlation is a measure of the strength of a relation (or association) rather than agreement between two factors. Due to the underlying statistical assumption of correlation coefficients, interpretation of test accuracy using these values can be misleading, as a correlation coefficient depends on the range of values in the sample, and the correlation is likely to be stronger when the range is wide rather than narrow. Furthermore, it is difficult to assign a correlation coefficient threshold value for widespread acceptance of a test. In this context, we focused on a diagnostic classification of patient LDL pattern type that goes beyond the quantitative data of particle size, as the latter parameter is often not comparable due to differences between the methodologies employed. Our results indicate that 94.4% of the study population had the same LDL particle pattern, with only 10 patients being misclassified. Despite this, there was very good agreement between the 2 methods, with a weighted kappa index of 0.81. Of the 32 samples classified by PGGE as sdLDL particles, 6 were misclassified by PTGE as large LDL particles (81.3% concordance), and of the 145 samples classified by PGGE as large, 4 were misclassified by PTGE as small (97.2% concordance). Consequently, this method of analysis appears to permit a simplified categorization of patients as displaying pattern A or pattern B, with good correlation and concordance with PGGE. To our knowledge, only three studies have addressed this topic previously, and all three classified LDL subfractions as small, intermediate or large based on different cut-off points. In accordance with our results, Hirany et al. [13] reported good agreement between PTGE and PGGE methods after evaluating the data using kappa statistics (weighed K =0.78), with the former method showing an agreement of 92%, 33% and 77% concordance for classification of small, intermediate and large LDL subfractions, respectively, when compared with PGGE. Similarly, another study reported 100%, 64%, and 95% agreement for classification of
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
A
255
B 28
LDL size diference (nm)
LDL diameter by PGGE (nm)
4 29 r=0.568, p<0.001
27 26 25 24 25
26
27
28
LDL diameter by PTGE (nm)
r = -0.617, p<0.001
3 2 1 0 -1 -2 24.5
25.5
26.5
27.5
Average LDL size (nm)
Fig. 2. LDL size analyzed by PGGE plotted against LDL size measured by PTGE (A) and Bland-Altman plot of absolute differences between LDL size determined by PTGE and PGGE vs the average of the two methods, with the mean of the differences (solid line) and mean ± 2SD limits (broken line) (B). Abbreviations: PTGE: polyacrylamide tube gel electrophoresis; PGGE: polyacrylamide gradient gel electrophoresis.
small, intermediate and large LDL subfractions, respectively, although the kappa index was not provided [12]. In a more recent systematic review, the abovementioned concordance percentages were re-examined, showing a noticeable change from 100% to 88% and from 95% to 24% for small and large LDL subfractions, respectively [11]. These differences were based on the assignment of subjects with an intermediate pattern. Finally, Ensign and coworkers [14] compared LDL pattern as determined by different technologies – PGGE, ultracentrifugation, nuclear magnetic resonance and PTGE – and demonstrated disparities among the methods, showing that complete agreement with respect to LDL subclass phenotype occurred in only 8% of the samples analyzed, reporting a mere 41% agreement between PTGE and PGGE with respect to the classification of LDL pattern (38% for large LDL and 3% for sdLDL). The small sample sizes of these studies [12,14] and/or their categorization of LDL in three patterns (the intermediate pattern being most prone to misclassification) could justify the variability in concordance indexes reported. Additionally, computer software for automated analysis of the bands was not available at the time the studies were published; therefore, each system would have had a different cut-off point for the classification of each pattern. Currently, the automated system of PTGE supplied by Quantimetrix (Lipoware software) allows a standardized analysis of LDL subfraction. In our Lipid Research Unit, we have observed that the PTGE method is precise and compares favourably with the PGGE method when employed for diagnostic classification of LDL particle type. Thus, we would affirm that both analytical methods can be used interchangeably. Additional advantages of PTGE with respect to PGGE is that it is faster and easier to perform when assessing large numbers of samples, and therefore could be of great practicality in clinical practice. However, PTGE is a relatively expensive procedure and may not be suitable for all patients. As a consequence, its use should be limited to patients in whom the determination of LDL particle size is likely to provide a real assessment of CAD risk. Table 2 Comparability of polyacrylamide tube gel electrophoresis and polyacrylamide gradient gel electrophoresis as methods of LDL subfraction determination. PGGE
Pattern A (LDL size ≥25.5 nm) Pattern B (LDL size b 25.5 nm)
Number of cases
PTGE Pattern A
Pattern B
(LDL size ≥26.8 nm)
(LDL size b26.8 nm)
Concordance (%)*
145
141
4
97.2
32
6
26
81.3
*weighed kappa of 0.81. Abbreviations: PTGE: polyacrylamide tube gel electrophoresis; PGGE: polyacrylamide gradient gel electrophoresis.
An increased proportion of sdLDL particles has been associated with marked alterations in serum lipoprotein and lipid levels, in particular elevated triglyceride concentrations. In accordance with that reported by other studies [12,13,20,23,24], we found that triglyceride and Apo B concentration –a indicator of the total number of LDL particles [25] – and AIP – considered a surrogate of sdLDL particle size [26]– were closely associated with LDL particle size. By dividing triglyceride levels into tertiles (firstb 68 mg/dl, second 68–116 mg/dl and thirdN 116 mg/dl), we were able to observe that the correlation of triglyceride and AIP with LDL particle size disappeared in the lower (first and second) tertiles, showing a good correlation, and one which was even better than in the general population, when only subjects in the third tertile were considered. Thus, triglyceride levels did not predict the LDL particle size of subjects in the first and second tertiles. Other lipoproteins may have been responsible for the variation of LDL particle size in our population, although this parameter was relatively constant, since only six of the 119 subjects was classified as pattern B (5.0%). Thus, Apo B showed a good correlation with LDL size (measured by PTGE) in all tertiles, which is in accordance with the results of a recent study that compared the PTGE classification method with measurement of Apo B and direct sdLDL [27]. In contrast, we detected a pattern that was predominantly of the B type (41.4%) and a more atherogenic lipid profile among patients in the third tertile that was associated with high total and LDL cholesterol (borderline high-risk), Apo B levels and AIP and low HDL cholesterol levels. In this way, our results suggest that determining LDL subfractions in patients with higher-than-normal triglyceride levels
Table 3 Correlation of triglycerides, atherogenic index and Apo B of plasma with LDL particle size measured by polyacrylamide gel gradient electrophoresis and polyacrylamide tube gradient electrophoresis. PTGE
Total triglycerides (mg/dl) 1st tertile 2nd tertile 3rd tertile AIP (log (triglycerides/HDLc) 1st tertile 2nd tertile 3rd tertile Apo B (mg/dl) 1st tertile 2nd tertile 3rd tertile
PGGE
r*
p
r*
p
− 0.705 0.086 − 0.123 − 0.676 − 0.607 0.055 − 0.123 − 0.636 − 0.532 − 0.305 − 0.565 − 0.361
b 0.001 0.522 0.339 b 0.001 b 0.001 0.686 0.342 b 0.001 b 0.001 0.021 b 0.001 0.005
− 0.484 0.129 − 0.028 − 0.527 − 0.435 0.094 0.088 − 0.552 − 0.244 0.051 − 0.321 0.060
b 0.001 0.338 0.830 b 0.001 b 0.001 0.487 0.496 b 0.001 0.001 0.705 0.012 0.652
*Pearson correlation. 1st tertile included 57 subjects with triglyceride levels b68 mg/dl, 2nd tertile included 62 subjects with triglyceride levels between 68 mg/dl and 116 mg/dl and 3 rd tertile included 58 subjects with triglyceride levels N 116 mg/dl.
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257
A LDL diameter (nm)
B
PTGE
27.5 a
a
27.3 27.1
b
26.9 26.7 26.5
<68
68-116
PGGE
26.5
LDL diameter (nm)
256
26.3
a
25.9 25.7 25.5
>116
a
26.1
b <68
Pattern A
Pattern B
Pattern B
58.6 100
90.3 414 41.4
0 <68
9.7 68-116
>116
Pattern A
100
50
68-116
Triglycerides (mg/dl)
>116
% of cases by PGGE
% of cases by PTGE
Triglycerides (mg/dl)
100
53.4 50
100
91.1
<68
8.1 68-116
66 0
Triglycerides (mg/dl)
>116
Triglycerides (mg/dl)
Fig. 3. Distribution of LDL diameter and pattern types measured by PTGE (A) and PGGE (B) according to triglyceride tertiles. Data are expressed as mean ± standard error of 57, 62 and 58 subjects for first, second and third tertile, respectively. Unlike superscript letters were significantly different when different tertiles were compared using one-way ANOVA and a Student-Newman-Keuls test as post hoc. Abbreviations: PTGE: polyacrylamide tube gel electrophoresis; PGGE: polyacrylamide gradient gel electrophoresis.
(N150 mg/dl) could be an important tool for assessing CAD risk, as this risk increases when values exceed 116 mg/dl. Considered as a whole, our results show that PTGE correlates favourably and is in good agreement with PGGE when used for the diagnostic classification of LDL pattern type. This evidence has led us to employ both analytical methods interchangeably in our Lipid Research Unit. Additionally, the determination of LDL particle size would seem to be an appropriate analytical procedure for estimating CAD risk in patients with borderline high-risk levels of total and LDL cholesterol and with triglyceride levels higher than 116 mg/dl. List of abbreviations AIP atherogenic index of plasma Apo B apolipoprotein B CAD coronary artery disease CHD coronary heart disease CVD cardiovascular disease FDA Food and Drug Administration K kappa index HDL high-density lipoprotein LDL low-density lipoprotein PGGE polyacrylamide gradient gel electrophoresis PTGE polyacrylamide tube gel electrophoresis Rf retardation factor sdLDL small dense low-density lipoprotein VLDL very-low density lipoprotein
Conflict of interest The authors claim no conflict of interest associated with the manuscript.
Acknowledgements Sources of financial support: The study has been supported by grants PS09/01025 and PI10/1195 from Fondo de Investigación Sanitaria (FIS) and AP-192/11 from Regional Ministry of Health of Valencian Community. M Rocha is a recipient of Miguel Servet contract from FIS (CP10/00360). VM Víctor is a recipient of Regional Ministry of Health of Valencian Community and Carlos III Health Institute contract (CES10/030). We kindly thank B Normanly and I Soria-Cuenca for their contribution to the present study.
References [1] Singh BK, Mehta JL. Management of dyslipidemia in the primary prevention of coronary heart disease. Curr Opin Cardiol 2002;17:503–11. [2] Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss RM. Lowdensity lipoprotein subclass patterns and risk of myocardial infarction. JAMA 1988;260:1917–21. [3] Krauss RM, Siri PW. Metabolic abnormalities: triglyceride and low-density lipoprotein. Endocrinol Metab Clin North Am 2004;33:405–15. [4] Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of The Third Report of The National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, And Treatment of High Blood Cholesterol In Adults (Adult Treatment Panel III). JAMA 2001;285:48–97. [5] Rizzo M, Berneis K. Low-density lipoprotein size and cardiovascular prevention. Eur J Intern Med 2006;17:77–80. [6] Otvos JD, Mora S, Shalaurova I, Greenland P, Mackey RH, Goff Jr DC. Clinical implications of discordance between low-density lipoprotein cholesterol and particle number. J Clin Lipidol 2011;5:105–13. [7] Hirano T, Ito Y, Yoshino G. Measurement of small dense low-density lipoprotein particles. J Atheroscler Thromb 2005;12:67–72. [8] Griffin BA, Caslake MJ, Yip B, Tait GW, Packard CJ, Shepherd J. Rapid isolation of low density lipoprotein (LDL) subfractions from plasma by density gradient ultracentrifugation. Atherosclerosis 1990;83:59–67. [9] Otvos JD, Jeyarajah EJ, Bennett DW. Quantification of plasma lipoproteins by proton nuclear magnetic resonance spectroscopy. Clin Chem 1991;37:377–86.
C. Bañuls et al. / Clinica Chimica Acta 413 (2012) 251–257 [10] Austin MA, Hokanson JE, Brunzell JD. Characterization of lowdensity lipoprotein subclasses: methodological approaches and clinical relevance. Curr Opin Lipidol 1994;5:395–403. [11] Chung M, Lichtenstein AH, Ip S, Lau J, Balk EM. Comparability of methods for LDL subfraction determination: A systematic review. Atherosclerosis 2009;205:342–8. [12] Hoefner DM, Hodel SD, O'Brien JF, et al. Development of a rapid, quantitative method for LDL subfractionation with use of the Quantimetrix Lipoprint LDL System. Clin Chem 2001;47:266–74. [13] Hirany SV, Othman Y, Kutscher P, Rainwater DL, Jialal I, Devaraj S. Comparison of low-density lipoprotein size by polyacrylamide tube gel electrophoresis and polyacrylamide gradient gel electrophoresis. Am J Clin Pathol 2003;119:439–45. [14] Ensign W, Hill N, Heward CB. Disparate LDL phenotypic classification among 4 different methods assessing LDL particle characteristics. Clin Chem 2006;52:1722–7. [15] Allain CC, Poon LS, Chan CS, Richmond W, Fu PC. Enzymatic determination of total serum cholesterol. Clin Chem 1974;20:470–5. [16] Bucolo G, David H. Quantitative determination of serum triglycerides by the use of enzymes. Clin Chem 1973;19:476–82. [17] Sugiuchi H, Uji Y, Okabe H, et al. Direct measurement of high-density lipoprotein cholesterol in serum with polyethylene glycol-modified enzymed and sulfates alphacyclodextrin. Clin Chem 1995;41:717–23. [18] Friedewald WT, Levy RI, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma without use of preparative ultracentrifuge. Clin Chem 1972;18:499–502. [19] Nichols AV, Krauss RM, Musliner TA. Nondenaturing polyacrylamide gradient gel electrophoresis. Methods Enzymol 1986;128:417–31.
257
[20] Kazumi T, Kawaguchi A, Hozumi T, et al. Low density lipoprotein particle diameter in young, nonobese, normolipidemic Japanese men. Atherosclerosis 1999;142: 113–9. [21] Altman DG. Some common problems in medical research. In: Altman DG, editor. Practical statistics for medical research. New York: Chapman and Hall; 1991. p. 396–437. [22] Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10. [23] Lamarche B, Tchernof A, Moorjani S, et al. Small, dense low-density lipoprotein particles as a predictor of the risk of ischemic heart disease in men. Prospective results from the Québec Cardiovascular Study. Circulation 1997;95:69–75. [24] Gardner CD, Fortmann SP, Krauss RM. Association of small low-density lipoprotein particles with the incidence of coronary artery disease in men and women. JAMA 1996;276:875–81. [25] Contois JH, McConnell JP, Sethi AA, et al. Apolipoprotein B and cardiovascular disease risk: Position statement from the AACC Lipoproteins and Vascular Diseases Division Working Group on Best Practices. Clin Chem 2009;55:407–19. [26] Dobiásová M, Frohlich J. The plasma parameter log (TG/HDL-C) as an atherogenic index: correlation with lipoprotein particle size and esterification rate in apoBlipoprotein-depleted plasma (FER(HDL)). Clin Biochem 2001;34:583–8. [27] Vandermeersch A, Ameye S, Puype D, Petitjean D, De Buyzere M, Langlois MR. Estimation of the low-density lipoprotein (LDL) subclass phenotype using a direct, automated assay of small dense LDL-cholesterol without sample pretreatment. Clin Chim Acta 2010;411:1361–6.