Disparate LDL phenotypic classification among 4 different methods assessing LDL particle characteristics.
Considerable heterogeneity exists in the physical properties (size and density) of LDL lipoprotein particles among groups of individuals (5). Moreover, many CVD patients do not have increased LDL-C (1, 6-8). These ambiguities were partially resolved by the demonstration of a significant association between LDL subclasses, intermediate-density lipoprotein, increased plasma triglyceride concentrations, and low plasma HDL-cholesterol (9). Adjustment for triglycerides, however, reduced the relative risk of myocardial infarction from 3 to 1.5 (95% confidence interval, 0.8-3.2), indicating a close association between certain types of LDL particle and plasma triglyceride concentration. Those investigators (9) grouped patients into type A and type B LDL patterns on the basis of the size/density distribution of the LDL particle subclasses (10-15).
There are now suggestions that environmental and metabolic factors can influence atherogenic profiles and potentially shift a type B LDL pattern toward a type A LDL pattern (16-18). Thus, clinical use of the LDL profile as a basis for dietary counseling, lifestyle adjustments, and the effectiveness of pharmacologic agents in reducing the risk of CVD could help guide potential treatment interventions (19).
Analytical ultracentrifugation is considered to be the benchmark for determining lipoprotein subfractions (20, 21). However, other methodologies have also been used, including nondenaturing gradient gel electrophoresis (GGE) (20,22), density gradient ultracentrifugation (23), and nuclear magnetic resonance (NMR).
Unfortunately, the estimated number of LDL subfractions is method dependent. Therefore, different methods may classify a patient's LDL particle type differently, making comparisons only tenuous and extrapolation to patient results difficult. The increasing clinical use of these newer technologies makes accurate and consistent patient classification by different laboratories crucial. To evaluate the performance of these new technologies for determining LDL subclasses and their effects on patient classification based on LDL subclass pattern, we conducted this comparison study.
Materials and Methods
Blood samples were collected into BD Vacutainer Tubes (BD Diagnostics) from 40 apparently healthy persons (30 males) 23 to 61 years of age. Serum and EDTA-plasma were kept at ambient temperature and separated within 2 h by centrifugation at 14308 for 15 min at 4[degrees]C, aliquoted, and then frozen at -70[degrees]C. Samples from each person were sent to 4 different laboratories for analysis. A portion of each serum sample was shipped frozen to Atherotech (Birmingham, AL), and another portion was retained frozen at our laboratory (Kronos Science Laboratories, Phoenix, AZ) for LDL subfractionation using the Lipoprint[TM] system (Quantimetrix). EDTA-plasma samples were shipped frozen for analysis to Liposcience (Raleigh, NC) and Berkley HeartLab (Alameda, CA). Each of the laboratories used a different method for LDL fractionation and LDL particle size determination (9).
METHOD 1: GGE
To assess LDL subfractions, Berkeley Heartlab uses LDL Segmented Gradient Gel Electrophoresis (LDL-[S.sub.3]GGE[TM]). This technique separates LDL particles into 7 LDL subfractions based on particle size and shape (20,22,23). The LDL-[S.sub.3]GGE for LDL typing classifications are as follows: 26.35-28.5 nm, large LDL (pattern type A); 25.75-26.34 nm, intermediate (pattern type AB); and 22.0-25.74 nm, small LDL (pattern type B). The LDL subfractions are reported as percentages based on the area under the curve for each of the 7 subfractions. Small LDL particles correspond to the subfractions LDL IIIa and LDL IIIb. These subfractions are indicators of the "severity" of the "atherogenic lipoprotein profile".
METHOD 2: ULTRACENTRIFUGATION-VERTICAL AUTO PROFILE
The Vertical Auto Profile-II, or VAP-II test (Atherotech), is derived from density gradient ultracentrifugation (24,25). The VAP-II generates a series of absorbance curves from which proprietary software produces 6 LDL subclasses labeled as LDL1 (most buoyant) through LDL6 (most dense). LDL1 and LDL2 comprise pattern type A and LDL3 and LDL4 comprise pattern type B (26).
METHOD 3: NMR
The laboratory report from Liposcience, Inc., contains 3 sections: a lipoprotein panel, a risk assessment panel, and a subclass panel. No references are provided for the basis of the "risk categories". The LDL size in nanometers and its classification as type A or B are also reported. The subclass panel contains the cholesterol content of the subclasses for intermediate-density lipoprotein and the triglyceride content for large, intermediate, and small VLDL (27).
METHOD 4: TUBE GEL ELECTROPHORESIS
Quantimetrix Lipoprint LDL System (28) was used according to the manufacturer's specifications at Kronos Science Laboratories, Inc. In this system, lipoprotein fractions are separated and summed in a weighted manner to yield an "LDLSF" score. Specific ranges of LDLSF scores correspond to different LDL phenotype patterns: normal (type A; LDLSF score <5.5), intermediate (type AB, LDLSF score 5.5-8.5), and atherogenic (type B; LDLSF score >8.5) (28).
This technology does not measure the LDL particle size directly, but estimates LDL particle size by comparing particle electrophoretic mobility with the electrophoretic mobilities of particles of known sizes. The total cholesterol of the sample must be measured independently of the Lipoprint system. The Lipoprint report contains the patient's scan and the cholesterol concentration within each of the fractions based on retardation factor ([R.sub.f]), which correspond to each of the LDL subfractions. The scan contains 7 possible LDL subfractions.
For across-method comparisons, we used a repeatedmeasures ANOVA and pairwise contrasts. Significance was established with the Bonferroni adjustment. When the sphericity assumption was in question, the Greenhouse-Geisler adjustment for the degrees of freedom was used to determine overall statistical significance. We used the intraclass correlation coefficient to establish the agreement among the within-subject data across laboratory values. When possible, the distribution of cholesterol contained in LDL and the estimates for LDL particle sizes for each laboratory were evaluated by use of boxplots. The cholesterol, triglyceride, and major lipoprotein cholesterol values were taken from the patient reports obtained from each method. Because this report primarily focuses on LDL particle characteristics and their use for LDL phenotype classification, HDL subfractions were not included in this analysis. All LDL particle sizes were converted to nanometers for comparisons across methods. A value of 5 nm was added to the NMR estimates of LDL particle size per the laboratory's statement included in the patient report. All statistical analyses were performed with the Statistical Package for the Social Sciences, Ver. 8.0 (SP55, Inc.).
The distribution of LDL phenotypes among the 40 persons for each method is shown in Fig. 1. Results varied considerably among the methods. The tube gel electrophoresis (TGE) classified 30 of 38 (79%) persons as type A pattern, whereas VAP classified only 3 of 39 (~8%) as type A. The VAP and NMR classified 21 persons (--~54%) as type B, but TGE classified only 2 of the 38 persons (~5%) as having a type B pattern. TGE and GGE classified 6 and 5 persons, respectively (n = 5) as having an intermediate pattern (type AB), whereas VAP classified 2.5 times as many persons (n = 15) as type AB. Although TGE and GGE classified the same number of persons as type AB, these methods did not agree on the AB classification for the same individuals. The laboratory using the NMR methodology does not include an intermediate (type AB) classification.
Patient classification data for each of the 4 methods are shown in Table 1. Complete agreement among the laboratory methods occurred in 3 (persons 2, 5, and 11) of 39 persons (~8%). NMR matched other methods to a greatest extent (marginal totals) and VAP matched the least. NMR and GGE produced the same result for 28 of 40 (70%) individuals. When we assumed that all AB patients reported by Berkeley Heart Laboratory (GGE) could reasonably be reported as pattern A, the agreement between NMR and GGE improved to 32 of 40 (80%). Twenty persons (~50%) were similarly classified by 3 of the 4 laboratories. Eight of these 20 individuals (~40%) were classified as type A by the TGE, NMR, and GGE methods, whereas 10 of 20 (50%) were classified type B by NMR, VAP, and GGE. Two of 20 persons (10%) were classified as type B when TGE was 1 of the 3 methods.
The original basis for distinguishing the LDL phenotype among individuals was LDL particle size (9). In 3 of the laboratory methods (TGE, GGE, and NMR), it was possible to determine the sizes of the main LDL particle fraction, providing the basis for LDL phenotypic classification, and in 1 case (GGE), 2 LDL particle sizes were given in the laboratory report, corresponding to LDL peaks 1 (GGE1) and 2 (GGE2). Total LDL-C could also be determined in 3 of the methods. Laboratory comparisons of LDL-C and particle size estimates are indicated in Fig. 2.
[FIGURE 1 OMITTED]
We observed significant differences among methods for LDL-C concentration (Fig. 2A). Although the absolute cholesterol concentrations were significantly different among methods, the within-subject LDL-C was relatively consistent (intraclass correlation = 0.904). These results indicate a degree of bias within each method, but the relative LDL-C within individuals was maintained across methodologies. The distribution of LDL particle sizes among the methods is shown in Fig. 2B. Differences among the methods for estimated LDL particle sizes were also significant. Unlike the LDL-C concentration, the within-subject particle size estimates were not sustained across methods (intraclass correlation = 0.246).
Several clinical laboratories, including the 2 largest national clinical reference laboratories, now offer LDL subfraction analysis performed with different technologies. The consistency of patient classification among laboratories is unknown because there is no standardization among the different methods. To gain insight into the consistency of LDL phenotype classification, we sent a sample from each of 40 patients to 4 different laboratories that use different methods for LDL subfraction analysis.
In this comparison study, we focused on the diagnostic classification of patient LDL pattern type. Typically, individuals with large LDL particles are classified as type A, whereas individuals with small, dense LDL particles are classified as type B and are at a greater risk for CVD (9,15). Our results indicated only 3 of 39 patients (~8%) were classified as having the same LDL phenotype by all 4 methods. The best agreement between 2 methods was between GGE and NMR (70%); this agreement may actually be >70%, as noted in the Results section. Removal of the 5 samples classified as AB, the 3 samples classified as not available (NA), and the 1 sample classified as indeterminate (IND) from the GGE report (Table 1), leaves 31 samples for comparison. The betweenlaboratory agreement for these 31 samples for the GGE and NMR methods would have been 90% (28 of 31).
The inability to consistently classify patients did not seem to be dependent on the LDL particle per se, because the inconsistency in classification was more or less evenly distributed among the methods used, suggesting that the classification is method specific. The absolute LDL-C concentration was significantly different among the compared methods, but the within-subject reproducibility (intraclass correlation = 0.904) was consistent across the methods. The LDL particle sizes were also significantly different among the methods, but unlike LDL-C, there was a poor correlation within individuals across methods. If the basis for patient classification is LDL particle size and if the particle size estimates for a given individual are different depending on the method used, then it is not surprising that patient classification across laboratories is not consistent.
As new methods for measuring LDL subclasses are introduced, standardization becomes increasingly important, particularly when these methods are based on dif ferent principles. If LDL particle size is to be meaningfully incorporated into a patient's heart disease risk profile, then a standardization program should be considered. At the very least, a proficiency testing program for routine interlaboratory comparison of test results should be implemented. Such a program could help identify and eliminate inconsistencies among methods. In the absence of such a program, the clinical utility of current methods for determining LDL subclasses phenotype is open to question.
[FIGURE 2 OMITTED]
Although our study demonstrated disparities among various methods for differentiating LDL particle size, certain potential limitations of our study cannot be ignored. Our samples were a combination of plasma and serum (provided as requested by each laboratory), and there may be a 3% to 7% difference in the concentration of lipids between plasma and serum (29, 30). Additionally, one cannot completely ignore the physical effects of freezing and thawing on LDL particle subclasses, but several studies have demonstrated that freezing and thawing do not alter LDL phenotypes when GGE or ultracentrifugation is used for analysis (31-33). Whether this lack of an effect holds true for the other methods must be determined by further investigation.
Received August 31, 2005; accepted May 4, 2006. Previously published online at DOI: 10.1373/clinchem.2005.059949
(1.) Griffin BA. Lipoprotein atherogenicity: an overview of current mechanisms. Proc Nutr Soc 1999;58:163-9.
(2.) Brown MS, Faust JR, Goldstein JL. Role of the low density lipoprotein receptor in regulating the content of free and esterified cholesterol in human fibroblasts. J Clin Invest 1975;55:783-93.
(3.) Goldstein JL, Ho YK, Basu SK, Brown MS. Binding site on macrophages that mediates uptake and degradation of acetylated low density lipoprotein, producing massive cholesterol deposition. Proc Natl Acad Sci U S A 1979;76:333-7.
(4.) Morel DW, DiCorleto PE, Chisolm GM. Endothelial and smooth muscle cells alter low density lipoprotein in vitro by free radical oxidation. Arteriosclerosis 1984;4:357-64.
(5.) Henriksen T, Mahoney EM, Steinberg D. Enhanced macrophage degradation of low density lipoprotein previously incubated with cultured endothelial cells: recognition by receptors for acetylated low density lipoproteins. Proc Natl Acad Sci U S A 1981;78: 6499-503.
(6.) Musliner TA, Krauss RM. Lipoprotein subspecies and risk of coronary disease [Review]. Clin Chem 1988;34:B78-83.
(7.) Fruchart JC, Packard CJ. Is cholesterol the major lipoprotein risk factor in coronary heart disease? A Franco-Scottish overview. Curr Med Res Opin 1997;13:603-16.
(8.) Castelli WP. Epidemiology of coronary heart disease: the Framingham study. Am J Med 1984;76:4-12.
(9.) Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC, Krauss RM. Low-density lipoprotein subclass patterns and risk of myocardial infarction. JAMA 1988;260:1917-21.
(10.) Campos H, Genest JJ Jr, Blijlevens E, McNamara JR, Jenner JL, Ordovas JM, et al. Low density lipoprotein particle size and coronary artery disease. Arterioscler Thromb 1992;12:187-95.
(11.) Austin MA, Hokanson JE, Brunzell JD. Characterization of lowdensity lipoprotein subclasses: methodologic approaches and clinical relevance. Curr Opin Lipidol 1994;5:395-403.
(12.) Frohlich JJ, Seccombe DW, Pritchard PH. The role of apoproteins in disorders of lipoprotein metabolism. Clin Biochem 1989;22: 51-6.
(13.) Slyper AH. Low-density lipoprotein density and atherosclerosis: unraveling the connection. JAMA 1994;272:305-8.
(14.) Krauss RM. Atherogenic lipoprotein phenotype and diet-gene interactions. J Nutr 2001;131:340S-3S.
(15.) Krauss RM, Siri PW. Metabolic abnormalities: triglyceride and low-density lipoprotein. Endocrinol Metab Clin North Am 2004;33: 405-15.
(16.) Ensign WY. Association of regional and total body composition, plasma lipids, lipoproteins, apolipoproteins and the effects of long-term resistance training in premenopausal females [PhD dissertation], AZ: University of Arizona, 1993.
(17.) Lamon-Fava S, McNamara JR, Farber HW, Hill NS, Schaefer EJ. Acute changes in lipid, lipoprotein, apolipoprotein, and low-density lipoprotein particle size after an endurance triathlon. Metabolism 1989;38:921-5.
(18.) Reaven GM, Chen YD, Jeppesen J, Maheux P, Krauss RM. Insulin resistance and hyperinsulinemia in individuals with small, dense low density lipoprotein particles. J Clin Invest 1993;92:141-6.
(19.) Shimabukuro M, Higa N, Asahi T, Oshiro Y, Takasu N. Fluvastatin improves endothelial dysfunction in overweight postmenopausal women through small dense low-density lipoprotein reduction. Metabolism 2004;53:733-9.
(20.) Krauss RM, Blanche PJ. Detection and quantitation of LDL subfractions. Curr Opin Lipidol 1992;3:377-83.
(21.) Krauss RM, Burke DJ. Identification of multiple subclasses of plasma low density lipoproteins in normal humans. J Lipid Res 1982; 23:97-104.
(22.) Nichols AV, Krauss RM, Musliner TA. Nondenaturing polyacrylamide gradient gel electrophoresis. Methods Enzymol 1986;128: 417-31.
(23.) Lee DM, Downs D. A quick and large-scale density gradient subfractionation method for low density lipoproteins. J Lipid Res 1982;23:14-27.
(24.) Kulkarni KR, Garber DW, Jones MK, Segrest JP. Identification and cholesterol quantification of low density lipoprotein subclasses in young adults by VAP-II methodology. J Lipid Res 1995;36:2291302.
(25.) Austin MA, King MC, Vranizan KM, Krauss RM. Atherogenic lipoprotein phenotype: a proposed genetic marker for coronary heart disease risk. Circulation 1990;82:495-506.
(26.) Chung BH, Segrest JP, Ray MJ, Brunzell JD, Hokanson JE, Krauss RM, et al. Single vertical spin density gradient ultracentrifugation. Methods Enzymol 1986;128:181-209.
(27.) Otvos JD, Jeyarajah EJ, Bennett DW, Krauss RM. Development of a proton nuclear magnetic resonance spectroscopic method for determining plasma lipoprotein concentrations and subspecies distributions from a single, rapid measurement. Clin Chem 1992; 38:1632-8.
(28.) Hoefner DM, Hodel SD, O'Brien JF, Branum EL, Sun D, Meissner I, 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.
(29.) Cooper GR, Myers GL, Smith SJ, Sampson EJ. Standardization of lipid, lipoprotein, and apolipoprotein measurements [Review]. Clin Chem 1988;34:B95-105.
(30.) Cloey T, Bachorik PS, Becker D, Finney C, Lowry D, Sigmund W. Reevaluation of serum-plasma differences in total cholesterol concentration. JAMA 1990;263:2788-9.
(31.) Dormans TP, Swinkels DW, de Graaf J, Hendriks JC, Stalenhoef AF, Demacker PN. Single-spin density-gradient ultracentrifugation vs gradient gel electrophoresis: two methods for detecting low density-lipoprotein heterogeneity compared. Clin Chem 1991;37: 853-8.
(32.) McNamara JR, Campos H, Ordovas JM, Wilson PW, Schaefer EJ. Gradientgel electrophoresis analysis of low density lipoprotein. Am Biotechnol Lab 1988;6:28-33.
(33.) Hokanson JE, Austin MA, Zambon A, Brunzell JD. Plasma triglyceride and LDL heterogeneity in familial combined hyperlipidemia. Arterioscler Thromb 1993;13:427-34.
WAYNE ENSIGN,  NICOLE HILL  and CHRISTOPHER B. HEWARD  *
 Naval Health Research Center, San Diego, CA.
 ICronos Science Laboratories, Inc., Phoenix, AZ.
 Nonstandard abbreviafions: CVD, cardiovascular disease; LDL-C, LDLcholesterol; GGE, gradient gel electrophoresis; NMR, nuclear magnefic resonance; VAP, ultracentrifugafion-verfical auto profile; and TGE, tube gel electrophoresis.
* Address correspondence to this author at: Kronos Science Laboratories, Inc., 2222 E. Highland Ave., Suite 220, Phoenix, AZ 85016. Fax 602-667-5623; e-mail Chris.firstname.lastname@example.org.
Table 1. Shows the patient classification data for each of the 4 laboratories along with results for total cholesterol, HDL-C, LDL-C, and triglycerides. Pattern type (a) TC, (b) HDL-C, LDL-C, TG, Patient TGE NMR VAP GGE mg/L mg/L mg/L mg/L 1 NA B AB B NA NA NA NA 2 A A A A 2090 450 1260 1400 3 A NA NA NA 1720 340 1220 510 4 A A B AB 2230 450 1610 1110 5 A A A A 2000 440 1290 650 6 AB B B B 1860 370 1110 1770 7 A B B B 2290 340 1240 3150 8 A A B AB 2110 490 1380 990 9 A B AB B 2260 350 1500 1370 10 AB B B B 1890 320 1170 2590 11 A A A A 2170 660 1310 760 12 AB B B B 2090 450 1280 1620 13 A A AB A 2150 500 1360 960 14 A B AB B 2320 460 1530 1680 15 A A AB NA 1980 350 1450 960 16 A B AB B 2530 340 1920 1410 17 AB B B B 2070 300 1460 2250 18 A A AB A 2250 380 1590 920 19 NA B B B NA NA NA NA 20 A B B B 2610 430 1730 2150 21 A A AB A 1640 380 1080 910 22 A B B A 1870 350 1200 1930 23 A B B A 1250 200 870 890 24 A A AB A 1900 360 1320 890 25 A A AB A 1170 330 630 830 26 A A B NA 2330 630 1380 1850 27 A B B IND 1360 270 8200 840 28 A A B A 1740 340 1190 1090 29 AB B B B 2460 460 1530 1510 30 A B AB B 2000 380 1190 1430 31 B B B A 1240 460 580 460 32 A A AB AB 1970 380 1230 1330 33 B B AB B 1820 330 710 4790 34 A A B A 2000 420 1280 1140 35 A A B A 2040 340 1390 1320 36 A B B B 1770 320 1100 1450 37 A B AB AB 1950 380 1220 1400 38 AB A AB A 1810 460 1170 370 39 A B B B 2830 450 1800 2620 40 A A B AB 1900 490 1200 490 (a) Shaded results do not match any other result for that patient. (b) TC, total cholesterol; HDL-C, HDL-cholesterol; TG, triglycerides; NA, test results not available; IND, indeterminate.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Lipids, Lipoproteins, and Cardiovascular Risk Factors|
|Author:||Ensign, Wayne; Hill, Nicole; Heward, Christopher B.|
|Article Type:||Clinical report|
|Date:||Sep 1, 2006|
|Previous Article:||Identification of novel brain biomarkers.|
|Next Article:||Sensitive gas chromatography-mass spectrometry method for simultaneous measurement of MDEA, MDMA, and metabolites HMA, MDA, and HMMA in human urine.|