1410

LDL Particle Size DistributionResults From the Framingham Offspring Study

Hannia Campos, Erling Blijlevens, Judith R. McNamara, Jose" M. Ordovas, Barbara M. Posner,

Peter W.F. Wilson, William P. Castelli, and Ernst J. Schaefer

Using 2-16% gradient gel electrophoresis, we examined low density lipoprotein (LDL) particle size inrelation to plasma lipoproteins in 1,168 women and 1,172 men from the Framingham Offspring Study. Inaddition, we studied the effect of dietary intake on LDL size in a subset of the population. Seven LDL sizepeaks were identified, with the largest, LDL 1, being found in the density range 1.019-1.033 g/ml; LDL 2and LDL 3 in d= 1.033-1.038 g/ml; LDL 4 and LDL S in d= 1.038-1.050 g/ml; and the smallest, LDL 6 and7, in d=1.050-1.063 g/ml. Seventy-seven percent of the population had one major and at least one minorLDL peak. Secondary LDL peaks accounted for 23% of the total LDL relative area, based on laserscanning densitometry. LDL size distribution was skewed toward larger LDL particles in women(prevalence of LDL 1, 30% and of LDL 2, 31%), whereas men exhibited a more symmetric distribution(prevalence of LDL 3, 42%). The prevalence of small (<255 A), dense (d>1.038 g/ml) LDL particles 4-7was 33% in men, 5% in premenopausal women, and 14% in postmenopausal women. In agreement withprevious reports, small, dense LDL particles were significantly (p<0.0001) associated with increasedtriglyceride and apolipoprotein (apo) B levels and decreased HDL cholesterol and apo A-I levels. Inaddition, we found a significant (/»< 0.0001) association between LDL cholesterol and LDL size. Thehighest LDL cholesterol levels were found among women with LDL 4 (148 mg/dl) and men with LDL 3-5(138 mg/dl). In addition, the presence of LDL 3 or 4 as secondary peaks was significantly associated withhigher LDL cholesterol levels, while smaller secondary LDL peaks were associated with higher triglyceridelevels. We also found that compared with subjects with optimal LDL cholesterol levels (<130 mg/dl),individuals with high-risk LDL cholesterol levels (a 160 mg/dl) had 1) a higher prevalence of LDL 3 and4 (women only) and a lower prevalence of LDL 1 and 2 (women only) and 2) 11% higher LDL cholesterolto apo B ratios, even when matched for LDL particle size. Furthermore, low saturated fat and cholesterolintakes were significantly associated (p<0.01) with smaller LDL particles. Therefore, the identification ofsmall, dense LDL particles per se may not be a good indicator of coronary artery disease risk in populationstudies. Gender differences and environmental factors that affect triglyceride levels and LDL physical andchemical properties should be taken into consideration. In addition, LDL 3-5 particles in men and 4 inwomen are associated with the highest LDL cholesterol levels. (Arteriosclerosis and Thrombosis1992;12:1410-1419)

KEY WORDS • LDL particle size • cholesterol • plasma lipoproteins • triglycerides • gradientgel electrophoresis • apolipoproteins • dietary fat intake

Low density lipoproteins (LDL) are the majorcholesterol-carrying lipoproteins in plasma.1

LDL isolated in the density region of 1.019-1.063 g/ml contain approximately (weight percent) 50%cholesterol (free and esterified), 25% protein, 20%phospholipid, and 5% triglyceride.1 Apolipoprotein

From the Lipid Metabolism Laboratory (H.C., E.B., J.R.M.,J.M.O., EJ.S.), US Department of Agriculture Human NutritionResearch Center on Aging at Tufts University, Boston, Mass.; theFramingham Heart Study (B.M.P., P.W.F.W, W.P.C.), NationalHeart, Lung, and Blood Institute, Framingham, Mass.; and theSchools of Medicine and Public Health (B.M.P.), Boston Univer-sity, Boston, Mass.

Supported by grant HL-35243; subcontract RFP NHLBIHV83-03 from the National Heart, Lung, and Blood Institute,National Institutes of Health, Bethesda, Md.; and contract 53-K06-5-10 from the US Department of Agriculture.

Address for correspondence: Dr. Ernst J. Schaefer, Lipid Me-tabolism Laboratory, USDA Human Nutrition Research Centeron Aging at Tufts University, 711 Washington Street, Boston, MA02111.

Received February 13, 1991; revision accepted August 4, 1992.

(apo) B, a 550-kd poh/peptide, comprises over 95% ofLDL protein mass.2 Elevated plasma levels of LDLcholesterol and apo B have been associated with pre-mature coronary artery disease (CAD).3"5

LDL particles are heterogeneous in size and density.6

On gradient gel electrophoresis, seven LDL subgroupscan be identified6-8 and have been shown to correlatewith specific LDL subclass density ranges.6-8-9 TheseLDL subclasses have been classified into two LDLphenotypes (patterns A and B).10 Several studies indi-cate that these two LDL phenotypes are inherited as asingle-gene trait with a dominant mode of inheri-tance.10-12 Other studies suggest that LDL subclassesare strongly influenced by environmental factors.13"18

The predominance of small (diameter <255 A), dense(d> 1.038 g/ml) LDL particles has been associated withthe presence of myocardial infarction (MI)19 andCAD,20-22 but this association is not independent oftriglyceride levels19-20 or established cardiovascular riskfactors, particularly LDL and high density lipoprotein(HDL) cholesterol levels.21-22

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

Campos et al LDL Particle Size Distribution 1411

It is well known that small, dense LDL particles areassociated with increased triglyceride and apo B levelsand decreased HDL cholesterol and apo A-I levels.8-912

These associations have been observed in several nor-mal populations,8916 in men with CAD,19-20'22 and infamilial combined hyperlipidemia.11 However, the asso-ciation between LDL subclasses and LDL cholesterollevels is unclear. While some studies indicate that thepredominance of small, dense LDL particles is associ-ated with significantly higher LDL cholesterol levels,8-12

other studies have reported either no association911 or atrend toward lower LDL cholesterol levels in subjectswith small, dense LDL.19-22 Furthermore, cross-culturalobservations indicate that while small, dense LDLparticles are more prevalent among populations withlower LDL cholesterol levels and lower dietary fatintake than those in the United States,16 these particlesare rarely found among study subjects from The Neth-erlands,9 a country where dietary fat consumption ishigh.23

In the present study we examined the distribution ofLDL subclasses, as well as the presence of secondaryLDL peaks, in relation to plasma lipoproteins, particu-larly LDL cholesterol levels, in 1,172 men and 1,168women from the Framingham Offspring Study. In addi-tion, we studied the association between dietary intakeand LDL particle size in a subset of this population.

MethodsStudy Subjects

The study subjects were 1,172 men and 1,168 womenin cycle 3 of the Framingham Offspring Study who hadcomplete data on LDL particle size and biochemicalparameters. Subjects were free of MI or CAD and werenot taking /3-blockers, oral contraceptives and/or estro-gen, or any other medications known to affect lipids. Allsubjects had blood drawn after a 12-hour fast as part ofan approved protocol as previously described.3

LDL Subclass DeterminationLDL subclasses were separated by subjecting whole

plasma to 2-16% gradient gel electrophoresis (PAA2-16%, Pharmacia, Piscataway, N.J.) and were visual-ized by using Sudan black to stain the LDL particles, aspreviously described.8 Scanning was performed on anLKB Ultrascan XL laser densitometer (LKB Instru-ments Inc., Paramus, N.J.) interfaced with an AT&Tcomputer (LKB) and a Canon PJ-108A printer usingthe LKB GSXL software for peak integration. Eachsubject was assigned an LDL type, with the largest,LDL 1, being found in the density range 1.019-1.033g/ml; LDL 2 and LDL 3 in the range 1.033-1.038 g/ml;LDL 4 and LDL 5 in the range 1.038-1.050 g/ml; andthe smallest, LDL 6 and 7, in the range 1.050-1.063g/ml. Since we previously found that 88% of subjectshave one major peak and one or two minor peaks, weestimated the percent relative area of each LDL peakafter scanning. The identification of each LDL peak wasbased on their relative distance from standard peaks ofa known plasma sample that was run in duplicate ineach gel.8 Both migration distance and the percent totalarea for each peak have been shown to remain constantthrough a large LDL concentration range and from gelto gel.8 Therefore, misclassification due to staining

LDLTYPE1LDL »COf» 1.44

LDLTYPE2LDL aeon 1*4

LDL1

LDL 2

LDL2

LDL1

56 44Xntatntro*

17 72 11

LDLtconUS ""-TYPEJ L D L t e < w , j . i 8

LDL 3

LOL2

LDL 3

LDL4

41 59 84 16

FIGURE 1. Top panel: Representative scans for women withpredominant low density lipoprotein (LDL) 1 and LDL 2 andLDL scores of 1.44 [(1x0.56)+ (2x0.44)] and 1.94[(1XO.17) + (2XO.72) + (3XO.11)J. Lower panel shows repre-sentative scans for two men with a predominant LDL 3 butwith different secondary peaks. The lower LDL score of 2.59[(2x0.41) + (3x0.59)] indicates the presence of larger LDLparticles compared with the higher LDL score of 3.16[(3x0.84)+4x0.16)].

and/or sample concentration is minimal. In this studywe found that 35% of the population had secondaryLDL peaks of at least 30% of the total area and that77% had one major and at least one minor secondaryLDL peak. To take into consideration the presence ofsecondary peaks, an LDL score for each subject wascalculated as the sum of the relative areas under allLDL peaks present. The boundaries for each LDL peakwithin the total LDL particle distribution were selectedat the points of inflection. The baseline was set atbackground on both sides of the LDL band range.Figure 1 (top panel) shows a representative scan for twowomen with predominant LDL 1 and LDL 2 and LDLparticle scores of 1.44 and 1.94 and for two men (lowerpanel) with a predominant LDL 3 and LDL scores of2.59 and 3.16. For example, for the first male subjectwith a predominant LDL 3 (59% of area) and asecondary LDL 2 (41% of area), an LDL score of 2.59was calculated [(3x0.59)+(2x0.41)]. For the secondmale subject with the same predominant LDL 3 (84% ofarea) but a smaller secondary LDL 4 (16%), an LDLscore of 3.16 was calculated [(3x0.84)+(4x0.16)]. Asmaller LDL particle score corresponds to a larger LDLparticle diameter.

Lipoprotein and Apolipoprotein AnalysesBlood was drawn from subjects after a 12-14-hour

fast in 0.15% EDTA (final concentration), and plasmasamples were centrifuged at 2,500 rpm for 20 minutes at4°C to isolate plasma. Plasma was subjected to ultracen-trifugation at <f=1.006 g/ml for 18 hours at 39,000 rpm,and the 1.006 g/ml supranatant and infranatant frac-tions were isolated. The HDL supernate was obtained

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

1412 Arteriosclerosis and Thrombosis Vol 12, No 12 December 1992

TABLE 1. LDL Subclass Distribution in WomenClassification as Previously Reportedf

LDLtype'

LDL

1 2

This study

peak area distribution (%)

3 4 5 6 7

and

n

F

Men From the Framingham Offspring Study and Corresponding LDL

(%) LDLMean subclasses

M LDL score (g/ml)t

LDLparticle

diameter

(A)*

Other studies

LDL LDLsubclass subtractionspattern§ (g/ml)||

Subgroup

LDL scoregroups#

1 64.6 24.5 9.9 0.6 0.2 0.1 0.1

2 9.2 74.9 13.4 1.8 0.2 0.2 0.2

3 1.6 6.6 76.2 14.3 0.7 0.2 0.2

4 0.2 0.9 14.7 73.4 10.0 0.2 0.2

5 0.3 0.2 5.1 10.6 80.7 2.0 1.1

6 0.0 0.2 0.3 3.8 2.3 85.7 7.7

7 0.0 0.0 0.0 0.0 12.5 6.0 81.5

348(30)364(31)345(29)47(4)53(5)9

(1)2

(0.2)1,168(100)

84(7)200(17)495(42)150(13)202(17)39(3)2

(0.2)1,172(100)

1.48±0.32

2.11+0.35

3.07+0.32

3.92±0.34

4.82+0.33

6.00+030

6.68+1.53

LDL I1.025-1.032LDLIIA1.030-1.038LDLIIB1.035-1.040LDLIIIA1.038-1.048LDLIIIB1.038-1.048LDLIVA1.048-1.065LDLIVB1.048-1.065

260-275

255-270

247-252

242-246

233-242

218-232

LDL-11.020-1.028LDL-21.027-1.034LDL-31.033-1.039LDL-41.039-1.049

LDL-51.049-1.061

Large

Large

Intermediate

Small

Small

Very small

LDL, low density lipoprotein.*As defined by McNamara et al.8

tAs defined by Nichols et al.7

tAs defined by Krauss.27

§As defined by Austin et al.10

||As defined by Swinkels et al.'#As defined by Campos et al.22

after precipitation of very low density lipoprotein andLDL with dextran-magnesium sulfate by using themethod of Warnick et al.24 Plasma total cholesterol,trigh/ceride, 1.006 g/ml infranatant cholesterol, andHDL cholesterol levels were determined enzymaticallywith an Abbott Diagnostics ABA-200 bichromatic ana-lyzer and Abbott A-GENT reagents.25 Plasma apo A-Iand apo B levels were determined with a noncompeti-tive enzyme-linked immunosorbent assay as previouslydescribed.26

Dietary Assessment and Statistical AnalysisTo examine the association between dietary intake

and LDL subclasses within a population, dietary infor-mation was obtained from a subsample of 85 womenand 76 men. Dietary intake was assessed by a self-administered food-frequency questionnaire as previ-ously described.16 The mean dietary intake for thispopulation has been previously reported.16 Statisticalanalyses were performed with the Statistical AnalysisSystems software (SAS, Cary, N.C.). The proceduresincluded t test analysis for mean comparisons of lipo-protein and apolipoprotein plasma parameters inwomen and men. The general linear model procedureand the least-square-means option were used for theLDL subclass analysis of variance, covariance, and ageadjustments. The LDL particle score distribution plotsand Pearson correlation coefficients were carried out byusing the Chart, Corr, Freq, and Univariate proceduresin the SAS system.

ResultsLDL Subclass Distribution in Women and Men

The population distribution for the seven LDL types,mean percent LDL peak area distribution, and mean

LDL particle score (LDL subclass classification, inwhich secondary LDL peaks are taken into consider-ation) for all women and men in this study, as well as thecorresponding LDL subclass classifications from previ-ous reports, are shown in Table 1. LDL 1 and 2 werefound most frequently among women. LDL 3 was themost frequently found LDL size among men. In men,33% had small LDL 4-7, and in women this percentagewas 10%. It should be pointed out that 45% Of thewomen in this population were postmenopausal, andthat menopausal status was significantly associated withLDL particle size distribution. The prevalence of smallLDL 4-7 was 5.3% in premenopausal compared with14.3% in postmenopausal women (^2<0.0001). Sincethis difference was no longer significant after adjustingfor age, we included all women as one group, and alldata were age adjusted.

Percent LDL peak area for each predominant LDLsize ranged between 65% and 86% of the total peakarea, while adjacent secondary peaks accounted for arange between 2% and 25% of the total peak area.Combinations among LDL subclasses that were morethan two LDL types apart (e.g., LDL 1 and LDL 4)were very rarely found (distant peaks accounted for<0.6% of area). The LDL score distribution for eachLDL type is shown in Figure 2. The LDL score popu-lations are normally distributed, with LDL 5 skewedtoward larger LDL particles. Since the calculated LDLscores are based on the percent peak area distribution,LDL scores that represent more than one LDL typeapart from the predominant LDL subclass (e.g., apredominant LDL 1 with and an LDL score ^3.0) wereuncommon (0.4% of the subjects).

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

Campos et al LDL Particle Size Distribution 1413

2 15 3 15 4 45

LDL TYPE 4n-1S7

3 15 4 U S 5i I B

LDL TYPE »n.SW

SO-

40-

30-

LDL TYPE 0n.2S4

1 II ! a J ]i I 4J

LOL TYPE 3n-941

3 15 4 4i 5 &6 6

LDL TYPEn-48

• 30

23 3 IS 4

LOL SCORE

IL8m«llK

STTUTTJ

' LDL SCORE

FIGURE 2. Bar graphs show the percent distribution of lowdensity lipoprotein (LDL) scores for six predominant LDLpeaks found among 1,168 women and 1,172 men from theFramingham Offspring Study.

To test whether the presence of secondary peaks wasassociated with lipoprotein parameters, we examinedassociations between LDL score and plasma triglycer-ide, HDL cholesterol, and LDL cholesterol levels byLDL type group, adjusted for age and gender, in womenand men. As shown in Table 2, secondary peaks insubjects with LDL 1 and LDL 2 were not associatedwith triglyceride or LDL cholesterol. In subjects withLDL 1, the presence of a smaller, secondary LDL peakwas significantly associated with higher HDL choles-terol levels. This LDL subclass profile was commonlyfound among women (see top panel of Figure 1 andFigure 3A). The presence of secondary LDL peaks wasmost informative for subjects with LDL type 3 (mostlyfound in men; see Figure 3B). In this group lower, large,

TABLE 2. Correlation Coefficients Between LDL Particle Scoreand UpWs by LDL Type

LDL type

LDL1LDL 2LDL 3LDL 4LDL 5LDL 6

n

43256484019725948

TG

-0.02-0.03

032*0.160.46*0.29'

HDL chol

0.13*-0.01-0.38$-0.02-0.06-0.26

LDL chol

-0.100.070.23*

-0.08-0.14*

-0.35f

LDL, low density lipoprotein; TG, triglyceride; chol, choles-terol. Analysis was carried out on data for all men and women(n=2340), with values adjusted for age and gender.

V < 0 . 0 5 , t 0 0

secondary peaks (e.g., LDL 1 or 2) and the presence ofsmaller secondary peaks (e.g., LDL 4; see lower panelof Figure 1) were associated with increased triglycerideand LDL cholesterol levels and decreased HDL choles-terol levels. A reduction in large, secondary LDL peakarea in subjects with LDL types 4, 5, and 6 was alsoassociated with increased triglyceride levels but not withHDL cholesterol levels. A reduction in large, secondaryLDL peaks in subjects with predominant LDL 4, 5, and6 was associated with lower LDL cholesterol levels. Itshould be noted that in this last group of subjects,"larger" LDL peaks corresponded mostly to large,secondary peaks of LDL 3 and 4 and not to LDL 1 and2 (see Table 1 for percent peak area distribution).Therefore, the presence of LDL 3 and 4 as secondarypeaks is associated with higher LDL cholesterol levels.

Figures 3A and 3B show the LDL score distributionin women and men. The LDL particle score distributionin women is skewed toward larger particles (skew-ness=1.0), as expected by a higher prevalence of womenwith predominant LDL 1 and LDL 2, while in men theLDL score distribution is more symmetric (skew-ness=0.3) because of a predominance of men with LDL3. The 10th, 25th, 75th, and 90th percentiles for LDLparticle scores are 1.35, 1.60, 3.00, and 3.60 in womenand 2.00, 2.60,4.00, and 5.00 in men, respectively. Thesepercentiles in men correspond with the LDL scorecutpoints that were previously used to identify large,intermediate, and small LDL groups22 (also see Table 1for comparison with other LDL subclass classificationsystems). Figure 3C shows the LDL distribution in allsubjects as well as the mean LDL scores for women andmen in this study. For comparison with previous reportsof LDL particle size in population studies, we haveprovided the mean LDL scores for men and womenfrom a healthy Costa Rican population who habituallyconsume a low-fat diet. Both women and men from thispopulation have been reported to have significantlysmaller LDL particles than do subjects from the UnitedStates.16 In addition, we have included the mean LDLscore previously reported for patients with CAD.22 Asshown in this figure, the mean LDL score in thesepatients is identical to that previously reported for theCosta Rican group.

LDL Subclasses and Plasma Triglyceride, HDLCholesterol, and Apolipoproteins

Mean plasma lipoprotein and apolipoprotein concen-trations, as well as mean LDL particle score and typefor women and men, are shown in Table 3. Men hadsignificantly (p< 0.0001) higher triglyceride, LDL cho-lesterol, and apo B levels and significantly lower HDLcholesterol and apo A-I levels compared with women.Mean LDL particle score and mean LDL particle typewere 3.33 and 3.26 for men and 2.42 and 2.26 forwomen, respectively (a higher value was obtained whenLDL scores were used instead of LDL type, indicatinga predominance of smaller rather than larger secondarypeaks when these are taken into consideration). Nosignificant differences in age and total cholesterol werefound between women and men.

The plasma parameters for the seven LDL types inwomen are shown in Table 4. By analysis of variance theassociations of all plasma parameters with LDL typewere significant (p<0.0001) after adjusting for age.

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

1414 Arteriosclerosis and Thrombosis Vol 12, No 12 December 1992

A 20

LDL particle score distribution

WOMEN

II

SU71

I1 2 3 4

LDL SCORE

S 7

B 20-1

JL h;ll

MENMM (n-1.172)

llll li .1..1 2 3 4 5 8 7

U ' f l t r LOL SCORE 8 m " " t r

TABLE 3. Mean Lipoprotein and Apolipoproteln Levels inWomen and Men From the Framingham Offspring Study

Parameter

Age (years)

Total cholesterol (mg/dl)

Triglyceride (mg/dl)

LDL cholesterol (mg/dl)

HDL cholesterol (mg/dl)

Apo B (mg/dl)

Apo A-I (mg/dl)

LDL particle score*

LDL typet

Women(n = l,168)

48.0±9.6

207±41

94±71

128±37

57±14

92±30

158±37

2.42±1.00

2.26±1.13

Men(" = 1,172)

47.7±10.0

209 ±38

134±102

134±34

45±12

107±33

136±33

3.33±1.14

3.26±1.24

P

0.4

0.1

0.0001

0.0001

0.0001

0.0001

0.0001

0.0001

0.0001

LDL, low density lipoprotein; HDL, high density lipoprotein;apo, apolipoprotein.

*LDL particle score for each subject was calculated as the sumof the relative areas under all the LDL bands present (see text fordetails).

tLDL type represents the predominant LDL peak observed ineach subject. A lower score or type represents a larger LDLparticle.

3.4* Coam Rk*nm>amimt.1t

'4.00 Cotta Pkmn mtn ni 164.02 UmwthCADnC22

3 4 S

LDL SCORE

6 7Smaller

FIGURE 3. Bar graphs show the percent low density lipopro-tein (LDL) particle size distribution in (panel A) women and(panel B) men from the Framingham Offspring Study whensecondary LDL peaks are taken into consideration (LDLscores). Arrows indicate the 10th, 25th, 75th, and 90thpercentiles. In women these percentiles correspond to LDLscores of 1.35, 1.60, 3.00, and 3.60 and in men of 2.00, 2.60,4.00, and 5.00, respectively. Panel C: Bar graph shows theLDL score distribution in all Framingham subjects(n=2,340). Arrows indicate the mean LDL score for men andwomen in this study, as well as previously reported LDL scoremeans for a healthy Costa Rican population who habituallyconsume a low fat-diet and for patients with coronary arterydisease.

Triglyceride levels were significantly lower in the large-LDL subclasses. Triglyceride levels increased with de-creasing LDL size, with the smallest, LDL 7, having amean triglyceride level of 632 mg/dl. However, only twosubjects were found in this last category. The differencein triglyceride levels between all LDL particle types wasstatistically significant (p<0.001). Apo B levels werealso lower in the large LDL particle types and increasedwith decreasing size. However, no significant differenceswere found within LDL types 4,5,6, and 7. These threegroups had the highest apo B levels (mean, 122 mg/dl).In contrast, HDL cholesterol and apo A-I levels de-creased with decreasing size. No significant differencesin HDL cholesterol were found between LDL scoregroups 4, 5, 6, and 7. These groups had the lowest HDL

cholesterol levels (mean, 39 mg/dl). Similar results wereobtained for apo A-I levels. The plasma parameters forthe seven LDL types in men shown in Table 5 aresimilar to the values observed in women, except forthose shown in previous tables, where there is a higherprevalence of men with smaller LDL particles. Bymultivariate analysis, triglyceride (42%) and HDL cho-lesterol (8%) levels accounted for 50% of the variancein LDL particle size in women and for 57% of thevariance in LDL size in men.

LDL Subclasses and LDL Cholesterol LevelsThe associations of LDL cholesterol and LDL type

were different from those observed for other lipopro-tein parameters (Tables 4 and 5). LDL cholesterollevels did not increase or decrease consistently withLDL particle size, as observed for triglyceride, apo B,HDL cholesterol, and apo A-I. The highest LDL cho-lesterol levels were found in women with LDL type 4(mean, 148 mg/dl) and in men with LDL types 3, 4, and5 (mean, 138 mg/dl). The lowest LDL cholesterol levelswere found in women with LDL 1, 2, and 7 (mean, 126mg/dl) and in men with LDL 1, 6, and 7 (mean, 118mg/dl). LDL cholesterol in an intermediate range wasfound in women with LDL 3,5, and 6 (mean, 132 mg/dl)and in men with LDL 2 (mean, 130 mg/dl). In sum, theassociations between LDL cholesterol and LDL sub-class distribution are similar in men and women, withthe higher LDL cholesterol levels being observed insubjects with intermediate and small LDL-sized parti-cles. The association between LDL cholesterol andLDL subclasses is parabolic, and it is not apparent whenPearson or Spearman correlations or linear regressionanalysis is used. In addition, when LDL subclasses aredivided in two groups, different associations betweenLDL cholesterol and LDL size are obtained, dependingon the sample selection procedures and the prevalenceof subjects at the extreme (large or small) LDLsubclasses.

To further explore the association between LDLcholesterol and LDL type, we compared LDL type

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

Campos et al LDL Particle Size Distribution 1415

TABLE 4. Plasma Upoprotein Levels by LDL Type in Women From the Framingham Offspring Study

Plasma parameter

Total triglycerideLDL cholesterolHDL cholesterolApoBApo A-I

LDL1(/i=348)

66±3124+2*67±184±1

171±2

LDL 2(II=364)

77±3128±2*58±188±1

159±2

LDL 3(n=345)

103±3

132±2t52±196±1

151±2

LDL 4("=47)

161±8148±544±2*

116±4*142±5*

LDL 5("=53)

212±8

133+5t41 ±2*

124+4'138±5*

LDL 6(n=9)

334±18130±12t33±4*

127±9*118±13*

LDL 7(«=2)

632±2975±1947±7.0*

141±15*155±20*

Analysis ofvanance p

<0.0001<0.0001<0.0001<0.0001<0.0001

LDL, low density lipoprotein; HDL, high density lipoprotein; apo, apolipoprotein. All parameters were adjusted for women aged 48 yearsand are given in milligrams per deciliter as mean±SEM. All groups are significantly different from each other (p<0.01) except as indicated(* or t), where groups with the same symbol are not significantly different from each other. By multivariate analysis triglyceride (/J2=0.42)and HDL cholesterol (/?2=0.08) accounted for 0.50 of the variability in LDL size. Apo B, LDL cholesterol, and apo A-I were alsoindependently associated with LDL size (p<0.01) and accounted for 0.02 of the variance in LDL size.

distribution in women and men with high-risk (^160mg/dl) and optimal (<130 mg/dl) LDL cholesterollevels, as based on the National Cholesterol EducationProgram cutpoints.28 As shown in Figure 4, women inthe high LDL cholesterol group displayed a significantlyhigher prevalence of LDL types 3 and 4 (33% and 11%,respectively) and a lower prevalence of LDL types 1 and2 (21% and 26%, respectively) relative to those in thelow LDL cholesterol group (26% and 2% for LDL 3 and4 and 35% and 33% for LDL 1 and 2, respectively).When compared with men in the low LDL cholesterolgroup, men in the high LDL cholesterol group had asignificantly higher prevalence of LDL type 3 (48%versus 37%, respectively) and a lower prevalence ofLDL type 1 (3% versus 11%, respectively).

Table 6 shows the LDL cholesterol to apo B ratio inwomen and men in each LDL type group and the highversus low LDL cholesterol groups. As shown in previ-ous studies, the LDL cholesterol to apo B ratio de-creased with increasing density in both LDL cholesterolgroups. It is important to point out that our apo B assaymeasures apo B within LDL and triglyceride-rich lipo-proteins. However, in all subjects most of the apo B inplasma is found within LDL. The highest LDL choles-terol to apo B ratio was found among subjects with highLDL cholesterol levels and large LDL 1 and LDL 2.When subjects with low and high LDL cholesterol levelswere matched for LDL type, those in the high LDLcholesterol group had, on average, an 11% higher LDLcholesterol to apo B ratio than those in the low LDLcholesterol group in all the LDL type groups. There-fore, subjects with high LDL cholesterol levels appearto have cholesterol-enriched LDL particles comparedwith those with low LDL cholesterol levels, despite their

having the same predominant LDL subclass. An in-creased prevalence of men with LDL 3 and women withLDL 3 and 4 was found in association with LDLcholesterol levels 2: 160 mg/dl.

LDL Subclasses and Dietary IntakeTable 7 shows the association between dietary intake

and LDL particle score in 161 men and women afteradjusting for caloric intake, age, and gender. Small,dense LDL particles were significantly associated withdecreased cholesterol intake (r=-0.27,/?<0.001). SmallLDL particles were also associated with decreased total,saturated, and monounsaturated dietary fatty acids(r=-0.18, p<0.05-0.01), as well as with decreased ani-mal fat consumption (r= —0.21,p<0.01). No associationbetween LDL particle score and protein, carbohydrate,or polyunsaturated fatty acid intake was found. Figure 5shows the mean cholesterol, animal fat, and saturatedfatty acid intake for the lower 25th, the 25th-75th, andthe upper 75th percentile groups, as well as the corre-sponding mean LDL score. Subjects in the lower 25thpercentile for dietary cholesterol (91 mg/1,000 kcal),animal fat (13% of calories), and saturated fat (9% ofcalories) had the smallest LDL particles (mean LDLscore±SEM, 3.29±0.08) in all the diet groups. Subjectsin the 25th-75th percentile for saturated fat intake(12.5% of calories) had significantly larger LDL particles(mean LDL score±SEM, 2.91±0.1) when comparedwith those in the lower 25th percentile group. The largestLDL particles (LDL score mean±SEM, 2.70+0.2) werefound among subjects in the upper 75th percentile forcholesterol intake (200 mg/1,000 kcal). Significantly(p<0.05) larger LDL particles were also found for

TABLE 5. Plasma Upoprotein Levels by LDL Type in Men From the Framingham Offspring Study

Plasma parameter

Total triglycerideLDL cholesterolHDL cholesterolApoBApo A-I

LDL1("=84)54±9

115±4*62+181±3

165±4

LDL 2(n=200)

78±5130±253±194±2

150±2

LDL 3(n=495)

1O4±3139±2f45±4

105 ±1135±1*

LDL 4(n = 150)

151±6140±3t42±1

116+2135+3*

LDL 5(n=202)

239±5136±2f37±1*

127±2*

122±2t

LDL 6("=39)

334±12116±5*33 ±2*

133±5*

118+51

LDL 7(«=2)

582±52124±23*37±7*

100±21*

126±21t

Analysis ofvariance p

<0.0001<0.0001<0.0001<0.0001<0.0001

LDL, low density lipoprotein; HDL, high density lipoprotein; apo, apolipoprotein. All parameters were adjusted for men aged 48 yearsand are given in milligrams per deciliter as mean+SEM. All groups are significantly different from each other (/?<0.01) except as indicated(* or t), where groups with the same symbol are not significantly different from each other. By multivariate analysis triglyceride (/?2=0.46)and HDL cholesterol (R2=0.U) accounted for 0.57 of the variability in LDL size. Apo B, LDL cholesterol, and apo A-I were alsoindependently associated with LDL size (p<0.01) and accounted for 0.02 of the variance in LDL size.

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

1416 Arteriosclerosis and Thrombosis Vol 12, No 12 December 1992

LDL<130mgftnl <n-M7)

UX>160 m|yral(n-21«)

X* -p<O0001

S *c£ 20Ha

10

HENLDL< 130 mtfrrl (H-S22)

LDL>160mgM (H-26G)

X -p<O0001

0 1Largar

3 4 SLDL TYPE

6 7SmalUr

FIGURE 4. Graphs show percentage of subjects with high-risk (^160 mg/dl) and optimal (<130 mg/dl) low densitylipoprotein (LDL) cholesterol levels for each predominantLDL peak.

subjects in the upper 75th percentile of animal fat intake(29% of calories).

DiscussionIn this study we examined LDL particle size distribu-

tion in 1,168 women and 1,172 men from the Framing-ham Offspring Study to establish normal LDL sizeranges and how these relate to diet and plasma lipopro-teins in a representative, randomly selected populationin the United States. In agreement with previous re-ports,8-29 our data indicate that women had a higherprevalence of LDL 1 and LDL 2 compared with men,who usually showed a predominance of LDL 3. Theprevalence of small, dense LDL was 33% in men. Thisfinding is consistent with other studies on LDL sub-classes among men of comparable age in the UnitedStates, where the prevalence of pattern B ranges be-tween 31% and 44%.1019 It has been reported thatpattern B is inherited as a single-gene trait with a

dominant mode of inheritance,10 which is fully ex-pressed in postmenopausal women with a frequency of49% compared with 13% in premenopausal women.10

Our study does not support this finding in women, sincethere were only 5% of premenopausal and 13% ofpostmenopausal women with a predominance of small,dense LDL particles in the Framingham population.

LDL subclasses have usually been analyzed in popu-lation studies by identification of the predominant LDLpeak via gradient gel electrophoresis.812-19 In our studywe examined the predominance of secondary LDLpeaks and their effect on plasma lipids. Our dataindicate that, overall, 77% of the total LDL peak area isaccounted for by the predominant LDL peak, 21% ofthe area is found in adjacent LDL secondary peaks, andless than 2% of the area is found in distant secondarypeaks. Furthermore, the presence of secondary LDLpeaks provided additional information with regard tothe associations between LDL particle size and LDLcholesterol and tnglycende levels. Our data indicatethat the presence of LDL 3 and LDL 4 as secondarypeaks is associated with higher LDL cholesterol levels,and the presence of smaller LDL peaks is associatedwith higher triglyceride levels. Additionally, in our studythe LDL particle size distribution, as calculated whensecondary LDL peaks were taken into consideration,was skewed toward larger LDL particles in women andwas more symmetrically distributed in men. Seventy-fivepercent of women had an LDL score ^3.0, and 75% ofmen had a value ^4.0. We did not find a bimodaldistribution of LDL particles, as previously reportedwhen the particle diameter of the predominant LDLpeak was used.30 These differences could be due to thesmaller sample size and the analysis of men and womenas one group in the previous study.30

It has been established that LDL particle size isassociated with increased triglyceride, VLDL mass,intermediate density lipoprotein mass, and apo B levelsand with decreased FTDL cholesterol, HDL2 choles-terol, and apo A-I concentrations.8912-27 Our presentdata are in agreement with these previous observations.HDL cholesterol and apo A-I levels decreased andtriglyceride and apo B levels increased consistently withdecreased LDL particle size. The association betweenLDL cholesterol and LDL particle size is more complex.While some studies indicate that higher LDL choles-terol levels are found among subjects with small, dense

TABLE 6. LDL Cholesterol to Apo B Ratio by LDL Type in Women and Men With Low (<130 mg/dl) and High (£160mg/dl) LDL Cholesterol Levels

LDL type

LDL1LDL 2LDL 3LDL 4LDL 5LDL 6

Women LDL cholesterol/apo B

LDL cholesterol<130 mg/dl

n

225

214

167

10

22

6

Ratio

1.46±0.021.45 ±0.02135±0.021.27±0.101.08±0.050.77±0.44

LDL cholesterol

n

46

56

72

25

16

3

£160 mg/dl

Ratio

1.63±O.O5t

1.60±0.05t1.48±0.04f1.36±0.071.20±0.062.11±0.60

Men LDL

LDL cholesterol

n

60

101

193

59

84

24

< 130 mg/dl

Ratio

1.39±0.03137±0.031.30±0.021.15±0.031.05±0.030.78 ±0.06

cholesterol/apo B

LDL cholesterol£160 mg/dl

n

7

35

126

42

46

8

Ratio

1.68±0.09t1.50 ±0.05*1.43 ±0.02t1.34 ±0.04*1.17 ±0.04*1.11 ±0.12*

LDL, low density lipoprotein; apo, apolipoprotein.Significantly different from low LDL cholesterol group at *p<0.05,

mean±SEM.tp<0.01, and 4p<0.001. Values are given as age-adjusted

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

Campos et al LDL Particle Size Distribution 1417

TABLE 7. Partial Correlation Coefficient Between DietaryIntake and LDL Particle Score

Nutrient (g) LDL particle score

ProteinCarbohydrateTotal fatSat FAMono FAPoly FAAnimal fatCholesterol

-0.120.10

-0.17'-0.18*-0.19f-0.01-0.21t-0.27$

LDL, low density lipoprotein; FA, fatty acids. Analysis wascarried out in a subset of 161 women and men, and values wereadjusted for age, gender, and caloric intake.

•p<0.05, 1p<0.0l, tp<0.00\.

LDL particles,8'12 other studies report either no associ-ation911 or a trend toward lower LDL cholesterol levelsin subjects with small, dense LDL.1922 Our data indicatethat the highest LDL cholesterol levels are found insubjects with LDL 3-5. Women with LDL 4 hadsignificantly higher LDL cholesterol levels (mean, 148mg/dl) than any other LDL type group, while thehighest LDL cholesterol levels in men were foundamong those with LDL 3, 4, and 5 (mean, 138 mg/dl).Lower mean LDL cholesterol levels were found amongsubjects with LDL 1, 2, 5 (in women), 6, and 7. Theparabolic association between LDL particle size andLDL cholesterol is of interest. In our view, human LDLparticles 3-5 may be the most common LDL sizeobserved in mildly hypercholesterolemic CAD patientsbecause these LDL sizes are associated with the highestLDL to HDL cholesterol ratios compared with controlsubjects (4.15 versus 3.48, respectively).22 In addition,LDL 1 and 2 were very rarely found among these malepatients.22 However, as shown in Figure 3, healthy CostaRican men have LDL particle scores identical to thosefound in patients with CAD. In the absence of anincreased LDL to HDL cholesterol ratio (2.78 in CostaRican men),16 these particles are less likely to be

Largtr

2.7 —

2J —

U —

3JJ _

3.1 —

312 _U _

8mU«r

CholutwiJ

ChotMtwol/t

SMurettdFA

28-75>h91 ±14

13 ±29 ± 1

14S±1S20±2

12£±1

200 ±2829±517±2

()Saturated FA (%)

FIGURE 5. Graph shows the mean low density lipoprotein(LDL) particle score for each dietary intake percentile. Tableshows the mean±SD intakes of cholesterol, animal fat, andsaturated fat (FA) in each percentile. *p<0.01, §p<0.05indicates significantly different from LDL particle score in thelower 25th percentile, with values adjusted for age and gender(n=161).

associated with atherosclerosis, provided that CADmortality in Costa Rica is significantly lower than in theUnited States.31

LDL particle composition studies and epidemiologi-cal observations have shown that small, dense LDL ischaracterized by increasing protein content and reduc-tions in cholesterol ester.8-32-33 In addition, it has beendemonstrated that LDL from patients with familialhypercholesterolemia contains more molecules of cho-lesterol ester than does LDL from normal individuals,even when the LDLs were matched for molecularweight.34 In our study we found that when subjects withhigh-risk (^160 mg/dl) and optimal (<130 mg/dl) LDLcholesterol levels were matched for LDL subclassgroups, those in the high LDL cholesterol group had, onaverage, 11% higher LDL cholesterol to apo B ratiosthan those in the low LDL cholesterol group. The LDLcholesterol to apo B ratio decreased with decreasingsize in both groups. Interestingly, women and men inthe high LDL cholesterol groups were characterized bya decreased prevalence of LDL 1 and 2 and an in-creased prevalence of LDL 3 and 4. LDL heterogeneityhas been associated with CAD in several stud-ies.19-22-34'35 While studies of monkeys have shown thatthe largest LDL particles are associated with atheroscle-rosis,35 either buoyant LDL I34-36 or small LDL IIIsubclasses are frequently found in patients withCAD.19-22 However, the association between LDL sizeand CAD is not independent of other established riskfactors, especially LDL and HDL cholesterol levels.22

Nevertheless, some properties of small, dense LDL perse, such as increased susceptibility to in vitro copper-induced oxidation, may increase their atherogenic po-tential.37 Additionally, the presence of small, denseLDL particles could be a marker of lipoprotein alter-ations that predispose to CAD.38

An important factor to consider when examiningLDL size as a risk factor for CAD in population studiesis the effect of dietary intake on LDL particle size.Experiments in nonhuman primates have indicated thatdiets high in saturated fat and cholesterol cause pro-gressive increases in LDL particle size because of anincrease in cholesterol ester content.39 These largeparticles are powerful predictors of atherosclerosis inmonkeys.40 In addition, isocaloric substitution of fish oilfor lard is associated with smaller, n-3 fatty acid-enriched LDL particles, fewer cholesterol ester mole-cules, and lower transition temperatures.41 Further-more, we have previously reported that populationscharacterized by the consumption of low-fat diets havea higher prevalence of small, dense LDL particles andlower LDL cholesterol levels than those found in theUnited States.16 Our data in this study support previousfindings on the association between dietary intake andLDL particle size. Lower saturated fat and cholesterolintakes are associated with smaller, denser LDL parti-cles within the Framingham population. Therefore,small LDL size per se is probably not a good indicator ofatherosclerotic risk in population studies. Other envi-ronmental factors that affect LDL size, such as dietaryintake, body habitus, physical activity, and use of med-ications, should be taken into consideration in popula-tion studies relating LDL size to CAD risk.13-1822

The precise biochemical factors regulating LDL par-ticle size are not completely understood. Clearly, an

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

1418 Arteriosclerosis and Thrombosis Vol 12, No 12 December 1992

important factor affecting LDL particle size is theplasma triglyceride level, as well as the efficiency of thelipolytic system as previously described.42-43 In our studytriglyceride levels aJone explained 44% of the variancein LDL particle size. Decreased carbohydrate contentof LDL protein and lipids, as well as a decreased LDLsialic acid content of apo B, has been associated withsmall LDL particles. These differences in the carbohy-drate content of LDL subspecies may be associated withdifferential production of such particles or variability intheir metabolic fate.44 In addition, unusual LDL sub-class profiles have been identified in patients withcholesteryl ester transfer protein deficiency, indicatingthat cholesteryl ester transfer protein activity may alterthe metabolism of small LDL.45 Lipoprotein and he-patic lipase activity may also be important determinantsof LDL heterogeneity within and between genders. Ithas been hypothesized that hepatic lipase is involved inthe conversion of intermediate density lipoprotein toLDL.46 In fact, hepatic lipase deficiency appears toprevent LDL formation, resulting in the accumulationof large, buoyant, LDL-like particles.47 Elevated hepaticlipase and decreased lipoprotein lipase activities havealso been associated with the predominance of small,dense LDL in normal subjects.48 Qearly, more needs tobe learned about the precise factors regulating LDLparticle subclasses.

In sum, LDL particle size distribution is skewedtoward larger particles in women and is more symmet-rical in men. The presence of smaller, secondary LDLpeaks is associated with increased triglyceride levels,while LDL 3 or LDL 4 as a secondary peak is associatedwith higher LDL cholesterol levels. The current dataare consistent with the concept that alterations intrigrycerides and HDL cholesterol are strikingly associ-ated with LDL particle size, so that individuals whohave the largest LDL particles also have the lowesttriglyceride and the highest HDL cholesterol levels. Inaddition, we report that individuals with the highestLDL cholesterol levels ( a 160 mg/dl) have a higherprevalence of LDL types 3 and 4, as opposed to LDL 1and 2, and an 11% higher LDL cholesterol to apo Bratio even when matched for LDL particle size. Fur-thermore, diets low in saturated fat and cholesterol areassociated with smaller LDL particles. Therefore, theidentification of small, dense LDL size per se may notbe a good indicator of CAD risk in population studies.Our data indicate that LDL particle sizes 3-5 areassociated with the highest LDL cholesterol levels.

References1. Mahley RW, Innerarity TL, Rail SC Jr, Weisgraber KH: Plasma

lipoproteins: Apolipoprotein structure and function. / Lipid Res1984;25:1277-1294

2. Yang CY, Chen SH, Gianturco SH, Bradley WA, Sparrow JT,Tanimura M, Li WH, Sparrow DA, DeLoof H, Rosseneu M, LeeFS, Gu ZW, Gotto AMJ, Chan L: Sequence, structure, receptor-binding domains and internal repeats of human apolipoproteinB-100. Nature 1986;323:738-742

3. Kannel WB, Feinleib M, McNamara PM, Garrison RJ, CastelliWP: An investigation of coronary heart disease in families: TheFramingham Offspring Study. Am J Epidemiol 1979;110:281-290

4. Castelli WP, Garrison RJ, Wilson PWF, Abbott RD, Kalousdian S,Kannel WB: Incidence of coronary heart disease and lipoproteincholesterol levels: The Framingham Heart Study. JAMA 1986^256:2835-2838

5. Brunzell JD, Sniderman AD, Albers JJ, Kwiterovich PO: Apopro-teins B and A-I and coronary artery disease in humans. Arterio-sclerosis 1984;4:79-83

6. Krauss RM, Burke DJ: Identification of multiple subclasses ofplasma low density lipoproteins in normal humans. / Lipid Res1982;23:97-104

7. Nichols AV, Krauss RM, Musliner TA: Nondenaturing poh/acryl-amide gradient gel electrophoresis. Methods Enzymol 1986;128:417-431

8. McNamara JR, Campos H, Ordovas JM, Peterson J, Wilson PWF,Schaefer EJ: Effect of gender, age, and lipid status on low densitylipoprotein subtraction distribution: Results of the FraminghamOffspring Study. Arteriosclerosis 1987;7:483-490

9. Swinkels DW, Demacker PNM, Hendricks JCM, van't Laar A:Low density lipoprotein subtractions and relationship of other riskfactors for coronary artery disease in healthy individuals. Arterio-sclerosis 1989;9:604-613

10. Austin MA, King MC, Vranizan KM, Newman B, Krauss RM:Inheritance of low-density lipoprotein subclass patterns: Results ofcomplex segregation analysis. Am J Hum Genet 1988;43:838-846

11. Austin MA, Brunzell JD, Fitch WL, Krauss RM: Inheritance oflow density lipoprotein subclass patterns in familial combinedhyperlipidemia. Arteriosclerosis 1990;10:520-530

12. Austin MA, King MC, Vranizan KM, Krauss RM: Atherogeniclipoprotein phenotype: A proposed genetic marker for coronaryheart disease risk. Circulation 1990;82:495-506

13. Williams PT, Krauss RM, Kindel-Jovce S, Dreon DM, VranizanKM, Wood PD: Relationship of dietary fat, protein, cholesterol,and fiber intake to atherogenic lipoproteins in men. Am J Clin Nutr1986;44:788-797

14. Lamon-Fava S, Fisher EC, Nelson ME, Evans WJ, Millar JS,Ordovas JM, Schaefer EJ: Effect of exercise and menstrual cyclestatus on plasma lipids, low density lipoprotein particle size, andapolipoproteins. / Chn Endocrinol Metab 1989;68:17-21

15. Lamon-Fava S, McNamara JR, Farber HW, Hill NS, Schaefer EJ:Acute changes in lipid, lipoprotein, and apolipoprotein, and low-density lipoprotein particle size after an endurance triathlon.Metabolism 1989^8:921-925

16. Campos H, Willett WC, Peterson RM, Siles X, Bailey SM, WilsonPWF, Posner BM, Ordovas JM, Schaefer EJ: Nutrient intakecomparisons between Framingham and rural and urban Puriscal,Costa Rica: Associations with lipoproteins, apolipoproteins, andlow density lipoprotein particle size. Arterioscler Thromb 1991; 11:1089-1099

17. Campos H, Bailey SM, Gussak LS, Siles X, Ordovas JM, SchaeferEJ: Relations of body habitus, fitness level, and cardiovascular riskfactors including lipoproteins and apolipoproteins in a rural andurban Costa Rican population. Arterioscler Thromb 1991;11:1077-1088

18. Lamon-Fava S, Jimdnez D, Christian JC, Fabsitz RR, Reed T,Carmelli D, Castelli WP, Ordovas JM, Wilson PWF, Schaefer EJ:The NHLBI Twin Study: Heritability of apolipoprotein A-I and B,and low density lipoprotein subclasses and concordance for lipo-protein (a). Atherosclerosis 1991;91:97-106

19. Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC,Krauss RM: Low-density lipoprotein subclass patterns and risk ofmyocardial infarction. JAMA 1988;260:1917-1921

20. Crouse JR, Parks JS, Schey HM, Kahl FR: Studies of low densitylipoprotein molecular weight in human beings with coronary arterydisease. J Lipid Res 1985;26:566-574

21. Swinkels DW, Demacker P, Hendriks J, Brenninkmeijer BJ, StuytP: The relevance of a protein-enriched low density lipoprotein asa risk for coronary heart disease in relation to other known riskfactors. Atherosclerosis 1989;77:59-67

22. Campos H, Genest J, Blijlevens E, McNamara JR, Jenner JL,Ordovas JM, Wilson PWF, Schaefer EJ: Low density lipoproteinparticle size and coronary artery disease. Arterioscler Thromb 1992;12:187-195

23. Keys A: Coronary heart disease in seven countries. Circulation1970;41(suppl I):I-1-I-198

24. Warnkk GR, Benderson JM, Albers JJ: Dextran sulfate-Mg^ pre-cipitation procedure for quantitation of high-density-lipoproteincholesterol. Clin Chem 1982^28:1379-1388

25. McNamara JR, Schaefer EJ: Automated enzymatic standardizedlipid analyses for plasma and lipoprotein fractions. Clin Chun Ada1987;166:l-8

26. Ordovas JM, Peterson JP, Santaniello P, Cohn JS, Wilson PWF,Schaefer EJ: Enzyme-linked immunosorbent assay for humanplasma apolipoprotein B. J Lipid Res 1987;28:1216-1224

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

Campos et al LDL Particle Size Distribution 1419

27. Krauss RM: Relationship of intermediate and low-density lipopro-tein subspecies to risk of coronary artery disease. Am Heart J1987;113:578-582

28. The Expert Panel: Report of the National Cholesterol EducationProgram expert panel on detection, evaluation, and treatment ofhigh blood cholesterol in adults. Arch Intern Med 1988;148:36-69

29. Krauss RM, Lindgren FT, Ray RM: Interrelationships amongsubgroups of serum lipoproteins in normal human subjects. ClinOwn Ada 1980;104:275-290

30. Krauss RM: The tangled web of coronary risk factors. Am Heart J1991;9O(suppl 2A)J6-41

31. Campos H, Mata L, Siles X, Vives M, Ordovas JM, Schaefer EJ:Prevalence of cardiovascular risk factors in rural and urbanPuriscal, Costa Rica. Circulation 1992;85:648-658

32. Teng B, Thompson GR, Sniderman AD, Forte TM, Krauss RM,Kwiterovich PO Jn Composition and distribution of low densitylipoprotein fractions in hyperapobetalipoproteinemia, normolipi-demia and familial hypercholesterolemia. Proc NatlAcad Sci U SA1983;80:6662-6666

33. Musliner TA, Giotas C, Krauss RM: Presence of multiple subpop-ulations of lipoproteins of intermediate density in normal subjects.Arteriosclerosis 1986;6:79-87

34. Patsch W, Ostlund R, Kuisk I, Levy R, Schonfeld G: Character-ization of lipoprotein in a kindred with familial hypercholesterol-emia. J Upid Res 1982;23:1196-1205

35. Rudel LL, Parks JS, Johnson FL, Babiak J: Low density lipopro-teins in atherosclerosis. J Upid Res 1986^27:465-474

36. Campos H, Roederer GO, Lussier-Cacan S, Davignon J, KraussRM: Predominance of large low density lipoprotein particles innormolipidemic patients with coronary artery disease, (abstract)Circulation 1991;84(suppl II):II-119

37. de Graaf J, Hak-Lemmers HLM, Hectors MPC, Demacker PNM,Hendriks JCM, Stalenhoef AFH: Enhanced susceptibility to invitro oxidation of the dense low density lipoprotein subtraction inhealthy subjects. Arteriosckr Thromb 1991;ll:298-3O6

38. Krauss RM: Low-density lipoprotein subclasses and risk of coro-nary artery disease. Curr Opin Lipidol 1991;2:248-252

39. Rudel LL, Pitts LL, Nelson CA: Characterization of plasma lowdensity lipoproteins on non-human primates fed dietary choles-terol. J Upid Res 1977;18:211-222

40. Rudel LL, Bond MG, Bullock BC: LDL heterogeneity and ath-erosclerosis in nonhuman primates. Ann N Y Acad Sci 1985;454:248-253

41. Parks JS, Gebre AK: Studies on the effect of dietary fish oil on thephysical and chemical properties of low density lipoproteins incynomolgus monkeys. / Upid Res 1991;32305-315

42. Eisenberg S, Gavish D, Oschry Y, Fainaru M, Deckelbaum RJ:Abnormalities in very low, low, and high density lipoproteins inhypertrigryceridemia: Reversal toward normal with bezafibratetreatment. / Chn Invest 1984;74:470-482

43. Deckelbaum RJ, Granot E, Oschry Y, Rose L, Eisenberg S:Plasma triglyceride determines structure-composition in low andhigh density lipoproteins. Arteriosclerosis 1984;4:225-231

44. LaBelle M, Krauss RM: Differences in carbohydrate content oflow density lipoproteins associated with low density lipoproteinsubclass patterns. / Upid Res 1990;31:1577-1588

45. Sakai N, Matsuzawa Y, Hirano K, Yamashita S, Nozaki S, UeyamaY, Kubo M, Tarui S: Detection of two species of low densitylipoprotein particles in cholesteryl ester transfer protein defi-ciency. Arterioscler Thromb 1991;ll:71-79

46. Rubinstein A, Gibson JC, Paterniti JR, Kakis G, Little A, GinsbergHN, Brown WV: Effect of heparin-induced liporysis on the distri-bution of apolipoprotein E in hepatic triglyceride lipase and lipo-protein lipase. J Clin Invest 1985;75:710-721

47. Auwerx JH, Marzetta CA, Hokanson JE, Brunzell JD: Large buoy-ant LDL-like particles in hepatic lipase deficiency. Arteriosclerosis1989;9:319-325

48. Campos H, Dreon DM, Krauss RM: Variations in hepatic lipaseand lipoprotein lipase associated with low density lipoprotein sub-class patterns, (abstract) Circulation 1991;84<suppl II):II-20

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from

and E J SchaeferH Campos, E Blijlevens, J R McNamara, J M Ordovas, B M Posner, P W Wilson, W P Castelli

LDL particle size distribution. Results from the Framingham Offspring Study.

Print ISSN: 1079-5642. Online ISSN: 1524-4636 Copyright © 1992 American Heart Association, Inc. All rights reserved.

Avenue, Dallas, TX 75231is published by the American Heart Association, 7272 GreenvilleArteriosclerosis, Thrombosis, and Vascular Biology

doi: 10.1161/01.ATV.12.12.14101992;12:1410-1419Arterioscler Thromb Vasc Biol. 

http://atvb.ahajournals.org/content/12/12/1410World Wide Web at:

The online version of this article, along with updated information and services, is located on the

  http://atvb.ahajournals.org//subscriptions/

at: is onlineArteriosclerosis, Thrombosis, and Vascular Biology Information about subscribing to Subscriptions:

  http://www.lww.com/reprints

Information about reprints can be found online at: Reprints: 

document.Permissions and Rights Question and AnswerFurther information about this process is available in theis being requested is located, click Request Permissions in the middle column of the Web page under Services.Clearance Center, not the Editorial Office. Once the online version of the published article for which permission

can be obtained via RightsLink, a service of the CopyrightArteriosclerosis, Thrombosis, and Vascular Biology Requests for permissions to reproduce figures, tables, or portions of articles originally published inPermissions:

by guest on May 28, 2018

http://atvb.ahajournals.org/D

ownloaded from