Original Research

Low-Density Lipoprotein Subclass Distribution Pattern

and Adiposity-Associated Dyslipidemia in
Postmenopausal Women
Kevin C. Maki, PhD, Michael H. Davidson, MD, Mary Sue Cyrowski, RD, Ann C. Maki, MS, RD, Phyllis Marx, MD
Chicago Center for Clinical Research, Chicago, Illinois
Key words: lipoproteins, hyperlipidemia, obesity, body fat distribution
Objective: A predominance of small, dense low-density lipoprotein (LDL) particles (subclass pattern B) is
associated with increased risk for coronary heart disease and is characterized by elevated triglycerides and
depressed high-density lipoprotein (HDL) cholesterol concentrations. The present analysis was undertaken to
assess the impact of LDL subclass distribution pattern and adiposity on serum lipids in postmenopausal women.
Methods: Anthropometric measurements and fasting lipid data were obtained from 254 postmenopausal
women 70 years of age or younger, not receiving sex hormone replacement, who were participating in a clinical
trial designed to assess the influence of hormone replacement regimens on coronary heart disease risk markers.
Results: The prevalence of LDL subclass pattern B was 32%. Triglyceride levels were higher and HDL
cholesterol lower (both p0.001) in women with pattern B vs. pattern A, but total and LDL cholesterol levels
did not differ. LDL subclass pattern contributed independently to the variance in HDL cholesterol (p0.001) and
log
e
triglyceride (p0.001) concentrations explained by anthropometric variables (waist circumference or body
mass index). Compared to women with LDL subclass pattern A and waist circumference below the median value
of 83.0 centimeters, those with pattern B and waist 83.0 centimeters had markedly lower HDL cholesterol
levels [44.0 (41.647.4) vs. 57.2 (54.160.3) mg/dL, mean (95% CI)] and increased triglyceride concentrations
[geometric mean 147.8 (131.6165.7) vs. 95.4 (88.2102.5) mg/dL].
Conclusions: These data suggest that adiposity and LDL subclass distribution pattern are independent
determinants of plasma triglyceride and HDL cholesterol concentrations in postmenopausal women.
INTRODUCTION
A predominance of small, dense low-density lipoprotein
(LDL) particles (LDL subclass pattern B) is associated with a
2- to 3-fold increase in risk for coronary heart disease [14].
LDL subclass pattern B is also characterized by several abnor-
malities of the plasma lipid profile, notably elevated triglycer-
ides and depressed high-density lipoprotein (HDL) cholesterol
concentrations [14] as well as other metabolic disturbances
including insulin resistance, glucose intolerance and a hyper-
coagulable state [57]. In addition, small, dense LDL particles
may have heightened atherogenicity due to greater susceptibil-
ity to oxidative modification and higher affinity for arterial wall
proteoglycans [8]. Thus, the increased risk of coronary heart
disease associated with the LDL pattern B phenotype may be
secondary to the atherogenic influence of small, dense LDL
particles, the cluster of metabolic disturbances which accom-
pany this phenotype or a combination of these factors [910].
Family studies have shown linkage of LDL particle size to
loci on chromosomes 6, 11, 16 and 19 [8]. One-third to one-half
of the variance in peak LDL particle diameter is explained by
genetic factors [1011]. Excess body fat, particularly abdom-
inal fat, is associated with LDL subclass pattern B, as well as
elevated triglycerides and depressed HDL cholesterol [1214].
Katzel and colleagues studied the interaction between LDL
subclass pattern and adiposity among 160 men [15]. Those with
Abbreviations: BMIbody mass index, EPATEating Pattern Assessment Tool, HDLhigh-density lipoprotein, LDLlow-density lipoprotein, Log
e
natural logarithm,
METmetabolic equivalents.
Funding for this research was provided by Novo Nordisk Pharmaceuticals, Inc., Princeton, NJ.
Address reprint requests to: Kevin C. Maki, PhD, Chicago Center for Clinical Research, 515 North State Street, 27
th
Floor, Chicago, Illinois 60610.
Journal of the American College of Nutrition, Vol. 19, No. 1, 2330 (2000)
Published by the American College of Nutrition
23
LDL subclass pattern B had higher triglycerides and lower
HDL cholesterol at any level of adiposity (percent body fat),
compared to those with LDL pattern A [15]. These data support
the concept that the genetic factors underlying LDL subclass
distribution amplify the unfavorable effects of obesity on tri-
glyceride and HDL cholesterol concentrations in men. This
may have implications for coronary heart disease risk, partic-
ularly in light of the rapidly growing body of evidence dem-
onstrating the importance of triglyceride-rich lipoproteins in
the atherogenic process [16].
The present analysis was designed to test the hypothesis that
postmenopausal women with LDL subclass pattern B would
have greater disturbances of the plasma lipid profile (higher
triglycerides and lower HDL cholesterol) than women with
LDL pattern A at any level of adiposity. A secondary objective
was to assess and compare the utility of body mass index and
waist circumference as indicators of adiposity-related dyslipi-
demia in postmenopausal women.
MATERIALS AND METHODS
The dataset used for these analyses consisted of information
collected at baseline from a group of 270 postmenopausal
women who participated in a clinical trial designed to assess
the influence of three hormone replacement regimens on cor-
onary heart disease risk markers. All subjects provided written
informed consent and the study protocol was approved by an
institutional review board (Schulman Associates, Cincinnati,
OH). An additional 229 women were screened but did not
qualify for participation.
Eligible women were less than 71 years of age with natural
or surgically-induced menopause at least 12 months prior to
randomization, confirmed by a plasma estradiol level 20
pg/mL. Exclusion criteria included use of hormone replacement
or lipid-altering agents within 10 weeks of the baseline plasma
lipid measurements. Also excluded were women whose body
mass index was 31.5 kg/m
2
, who were heavy smokers (20
cigarettes per day) or alcohol users (14 alcoholic drinks per
week) or who engaged in substance abuse. Women with un-
controlled hypertension (systolic pressure 160 mm Hg, or
diastolic pressure 95 mm Hg) or elevated triglycerides (350
mg/dL at two consecutive visits) were excluded. Other medical
conditions excluding participation were history of stroke, pan-
creatitis, gallbladder disease, thrombophlebitis or thromboem-
bolic disorders, myocardial infarction within six months, ab-
normal genital bleeding of unknown etiology, an abnormal
mammogram suspicious for malignancy, the presence of he-
patic enzymes more than twice the upper limit of normal,
diabetes mellitus or other endocrine disease (except hypothy-
roidism adequately treated with a stable dose of thyroid re-
placement), and significant psychiatric disorders. Women using
beta-adrenergic blockers, high doses of thiazide diuretics (25
mg/d of hydrochlorothiazide or its equivalent), erythromycin,
immunosuppressants, systemic corticosteroids or anticoagu-
lants were also excluded.
Blood for baseline biochemical measures, including a
plasma lipid profile, glucose, insulin, and hemoglobin A
1C
, was
collected after an overnight fast at two baseline visits, approx-
imately 14 days apart. The mean of two values obtained on
separate days was used in the analyses for all biochemical
variables except LDL subclass distribution pattern and hemo-
globin A
1C
, which were measured once at the final baseline
visit. Plasma lipid profiles included total cholesterol, HDL
cholesterol, triglycerides and calculated values for LDL cho-
lesterol. LDL subclass pattern was determined once from a
plasma sample obtained at the final baseline visit.
Biochemical Analyses
Except for LDL subclass distribution, all biochemical as-
says were completed by Quest Nichols Institute, San Juan
Capistrano, CA. Quest Nichols Institute participates in the
Centers for Disease Control and Prevention/National Heart,
Lung, and Blood Institute lipid measurement standardization
program. LDL subclass distribution measurements were com-
pleted by Atherotech, Inc., Birmingham, AL.
The Vertical Auto Profile II method was used to assess the
concentration of cholesterol carried in large, buoyant (LDL
1
and LDL
2
) and small, dense (LDL
3
and LDL
4
) LDL particles,
as described elsewhere in detail [1718]. Briefly, the Vertical
Auto Profile II method utilizes single vertical spin density
gradient ultracentrifugation to separate the various plasma li-
poprotein fractions. After centrifugation, the cholesterol con-
tent of the tube is continuously analyzed and digitized. A
cholesterol absorbance curve profile is generated by plotting
digitized absorbance units on the Y axis and the relative gra-
dient position on the X axis. A deconvolution program is used
to separate the different lipoprotein classes and subclasses.
Subjects with 50% of their LDL cholesterol in the small,
dense fractions (LDL
3
LDL
4
) were classified as having the
small, dense LDL phenotype (LDL subclass pattern B).
Plasma cholesterol, triglyceride and glucose concentrations
were determined with a Hitachi 914 analyzer (Boehringer
Mannheim, Indianapolis, Indiana) which employs enzymatic
methods. HDL cholesterol was quantified after precipitation of
lower-density lipoproteins with phosphotungstate and magne-
sium. LDL cholesterol in mg/dL was calculated using the
following equation: LDL cholesteroltotal cholesterolHDL
cholesteroltriglycerides/6.25 [19]. This equation loses accu-
racy when the plasma triglyceride level exceeds 400 mg/dL.
Accordingly, no LDL cholesterol value was calculated in cases
where triglycerides were above this level. Hemoglobin A
1C
was measured with a VARIANT Analyzer (Bio-Rad Labora-
tories, Hercules, CA) by ion exchange high performance liquid
chromatography. Plasma insulin concentration was assessed by
radioimmunoassay (Linco Scientific, St. Charles, MO).
LDL Subclass Pattern B
24 VOL. 19, NO. 1
Questionnaires
Subjects completed a standard medical history question-
naire that was used to identify possible exclusion criteria and to
assess smoking and alcohol consumption habits. The Stanford
Seven-Day Physical Activity Recall questionnaire was used to
estimate energy expended during sleep, light, moderate, hard
and very hard activities [20]. Hours of activity in each category
were multiplied by constants to produce estimates of energy
expenditure. Estimated energy expenditure from each of these
categories was then summed to produce a physical activity
score in metabolic equivalent-hours per week (one metabolic
equivalent-hour represents approximately one kilocalorie per
kilogram of body weight). Dietary intake was assessed with
section one of the Eating Pattern Assessment Tool that
consists of questions relating to intake of foods in 11 categories
[21]. Lower scores indicate lower consumption of foods high in
fat, saturated fats and cholesterol. A score of approximately 28
or below is consistent with the dietary recommendations of the
National Cholesterol Education Program.
Anthropometric Measurements
Body weight and height were measured in light clothes
without shoes. Body mass index was calculated as weight in
kilograms divided by squared height in meters. Waist was
measured in duplicate at the minimum circumference between
the lowest rib and the iliac crest. If values differed by more than
0.5 cm, a third measurement was obtained, and the two closest
values were averaged.
Statistical Methods
Statistical analyses were completed using the Statview 4.5
(Abacus Concepts, Berkeley, CA) and JMP 3.1 (SAS Institute,
Cary NC) software packages. Plasma insulin, triglycerides and
physical activity score were not normally distributed. Natural
logarithm transformations produced acceptable distributions
for insulin and triglycerides, but not physical activity score.
Accordingly, physical activity score was ranked, and the ranks
were used in multivariate analyses. Analysis of variance,
Mann-Whitney U and Pearson chi-square tests were employed
to assess differences in characteristics of subjects with LDL
subclass patterns A and B.
Least squares linear regression models were fit for log
e
triglyceride, HDL cholesterol and LDL cholesterol in order to
test the null hypothesis that the regression lines for waist
circumference and body mass index on plasma lipid levels were
coincident for women with LDL subclass patterns A and B
[22]. A single regression model approach was used as described
by Kleinbaum and colleagues [22]:
y
1
x
1

2
x
2

3
x
1
x
2
error
where y is the lipid variable under investigation, x
1
is an
anthropometric variable (waist or body mass index) and x
2
is
LDL subclass pattern (0 A, 1 B). If the coincident lines
hypothesis was rejected, additional tests were run to assess
possible differences in slopes and intercepts. F-ratios calculated
for these tests used the mean squared error from the full model
as the denominator [22]. Separate regression models were also
fit for women with the two LDL subclass patterns. Correlation
coefficients are reported to express the strength of the relation-
ship between anthropometric measures and plasma lipid vari-
ables within LDL subclass categories. Analysis of variance was
employed to assess the influence of adiposity and LDL subclass
distribution pattern on mean serum lipid concentrations using a
median split to classify women into high and low catego-
ries for waist and body mass index.
The investigators felt that the deconvolution model em-
ployed to assess LDL subclasses provided a poor fit to the
observed data for 12 subjects. Separate analyses were com-
pleted for which these women were excluded. Since doing so
did not materially alter the results, only data from the full study
sample are presented.
RESULTS
LDL subclass distribution was not measured for 12 of the
270 women randomized because an inadequate volume of
plasma was available. An additional four women were ex-
cluded from the analyses because data for height were unavail-
able. Therefore, the analyses presented herein represent data
from 254 subjects.
Characteristics of the study sample categorized by LDL
subclass pattern are shown in Table 1. The prevalence of LDL
subclass pattern B was 32%. Women with pattern B were
slightly, but not significantly, older than those with pattern A.
Dietary fat intake, as indicated by the Eating Pattern Assess-
ment Tool, alcohol consumption and prevalence of current
cigarette smoking did not differ between LDL subclass groups.
Body mass index (p0.037) and waist circumference
(p0.031) were significantly higher among women with LDL
pattern B, while physical activity score was lower (p0.007).
The race/ethnicity of subjects in both LDL subclass categories
was predominantly caucasian (non-Hispanic white). Use of
antihypertensive medication and history of atherosclerotic dis-
ease were infrequent in both groups (8%). Differences were
not significant, but the prevalence of these characteristics tended to
be higher among subjects with LDL subclass pattern A.
Table 2 summarizes the biochemical characteristics of the
participants grouped by LDL subclass pattern. Women with
LDL subclass pattern B did not differ from pattern A subjects
with regard to total cholesterol, non-HDL or LDL cholesterol
levels. However, women with pattern B had marked elevations
in the concentration of cholesterol carried in the small, dense LDL
fractions (LDL
3
LDL
4
, p0.001), with proportionately less car-
ried in the larger, more buoyant fractions (LDL
1
LDL
2
,
LDL Subclass Pattern B
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION 25
p0.001). Women with LDL pattern B also showed the other
lipid abnormalities which characterize this phenotype, including
depressed HDL cholesterol, elevated triglycerides and increased
total/HDL cholesterol ratio (all p0.0001). Fasting plasma glu-
cose (p0.031), insulin (p0.002) and hemoglobin A
1C
(p0.016) levels were also significantly higher among those with
LDL subclass pattern B.
The two anthropometric indicators used to assess adiposity,
waist and body mass index, were significantly correlated in this
sample (r0.77, p0.001). The null hypothesis of coincident
regression lines was not rejected for waist or body mass index
in relation to LDL cholesterol, but was rejected (p0.001) for
both anthropometric measures in relation to HDL cholesterol
and log
e
triglycerides (Table 3). For these lipid parameters,
intercepts of the regression lines were significantly different
between LDL subclass patterns A and B (p0.001). The re-
gression lines for waist and body mass index did not differ
significantly in slope between the two LDL subclass groups.
Nevertheless, for both anthropometric measures, clear trends
were present toward steeper slopes among women with LDL
subclass pattern B, with p-values for the non-parallelism test
ranging from 0.09 to 0.24. The relationships between waist
circumference and log
e
serum triglyceride and HDL cholesterol
concentrations according to LDL subclass distribution pattern
are shown graphically in Fig. 1 and 2.
Table 1. Characteristics of the Study Sample according to
Low-Density Lipoprotein Subclass Pattern
Variable
LDL Subclass
Pattern A
n173
LDL Subclass
Pattern B
n81
p
Value
Age
1
58.9 (6.3) 60.2 (5.9) 0.118
Body mass index
1
,
(kg/m
2
) 25.7 (3.3) 26.6 (2.8) 0.037
Waist
1
, (cm) 82.3 (9.6) 85.1 (9.2) 0.031
EPAT score (part 1)
1
22.8 (5.1) 22.2 (4.9) 0.347
Alcohol intake
2
,
ounces/week
1.0 0.3 0.478
(0.0, 3.0) (0.0, 3.0)
Physical Activity
2
,
(MET-hour/week)
279.8 268.8 0.007
(257.0, 310.8) (249.1, 291.3)
Caucasian, % 83.8 87.0 0.837
Current smoker, % 14.2 13.8 0.924
Antihypertensive
medication use, % 7.4 5.0 0.556
History of
atherosclerotic
disease, % 6.9 2.5 0.154
Abbreviations: EPATEating Pattern Assessment Tool; LDLlow-density li-
poprotein; METmetabolic equivalents.
1
Values are mean (SD).
2
Values are median (25
th
, 75
th
percentile).
Table 2. Biochemical Characteristics of the Study Sample
according to Low-Density Lipoprotein Subclass Pattern
Variable
LDL Subclass
Pattern A
n173
LDL Subclass
Pattern B
n81
p
Value
Total cholesterol
1
(mg/dL) 228.5 (38.3) 224.9 (43.4) 0.495
Non-HDL cholesterol
1
(mg/dL)
172.9 (40.0) 177.7 (44.1) 0.381
LDL cholesterol
1
(mg/dL) 155.0 (36.3) 152.9 (38.3) 0.674
LDL
1
LDL
2
cholesterol
1
(mg/dL)
74.0 (19.0) 39.2 (15.7) 0.001
LDL
3
LDL
4
cholesterol
1
(mg/dL)
37.9 (15.9) 71.6 (19.6) 0.001
HDL cholesterol
1
(mg/dL) 55.7 (14.2) 47.1 (12.5) 0.001
Triglycerides
2
(mg/dL) 98.9 133.8 0.001
(93.3, 104.6) (121.0, 147.8)
Total/HDL cholesterol
ratio
1
4.35 (1.32) 5.07 (1.60) 0.001
Fasting plasma glucose
1
(mg/dL)
92.7 (9.3) 96.6 (11.1) 0.031
Fasting plasma insulin
2
(mU/L)
12.2 14.3 0.002
(11.6, 12.9) (13.0, 15.6)
HbA
1c
1
, % 5.59 (0.47) 5.74 (0.51) 0.016
Abbreviations: HbA
1C
hemoglobin A
1C
; HDLhigh-density lipoprotein;
LDLlow-density lipoprotein.
1
Values are mean (SD).
2
Values are geometric mean (95% confidence interval).
Table 3. Results of Least Squares Linear Regression
Analyses Showing the Relationships between Anthropometric
Indicators and Plasma Lipid Variables according to Low-
Density Lipoprotein Subclass Pattern
Independent Variable
and LDL Subclass
Pattern
Intercept Slope Pearson r
p
Value
Dependent VariableLDL Cholesterol
Concentration (mg/dL)
Waist (cm)
Pattern A 159.7 0.057 0.015 0.845
Pattern B 120.3 0.384 0.091 0.427
BMI (kg/m
2
)
Pattern A 142.6 0.473 0.043 0.574
Pattern B 156.7 0.142 0.010 0.926
Dependent VariableHDL Cholesterol
Concentration (mg/dL)
Waist (cm)
Pattern A 80.3 0.300 0.202 0.008
Pattern B 98.4* 0.602 0.439 0.001
BMI (kg/m
2
)
Pattern A 82.5 1.037 0.244 0.001
Pattern B 99.5* 1.968 0.448 0.001
Dependent VariableLog
e
Triglyceride
Concentration (mg/dL)
Waist (cm)
Pattern A 4.19 0.005 0.124 0.107
Pattern B 3.47* 0.017 0.335 0.003
BMI (kg/m
2
)
Pattern A 4.04 0.021 0.192 0.012
Pattern B 3.77* 0.042 0.264 0.017
Abbreviations: BMIbody mass index; HDLhigh-density lipoprotein;
LDLlow-density lipoprotein; log
e
natural logarithm.
* Significantly different from value for those with LDL subclass pattern A
(p0.001).
LDL Subclass Pattern B
26 VOL. 19, NO. 1
LDL cholesterol did not correlate significantly with anthro-
pometric indicators of adiposity within either LDL subclass
category (p values0.40). HDL cholesterol concentration was
significantly inversely correlated with waist girth and body
mass index within both LDL subclass groups (p0.01). Sig-
nificant positive associations were present for waist and body
mass index with log
e
triglyceride concentration among women
with LDL subclass pattern B (p0.02). Among women with
LDL subclass pattern A, log
e
triglyceride concentration was
associated with body mass index (p0.02), but the association
did not reach the 5% level of significance for waist circumfer-
ence (p0.107, p0.10).
Waist circumference alone explained 8.6% of the variance
in HDL cholesterol and 5.2% of the variance in log
e
triglycer-
ide concentration (p0.001 for both). The addition of LDL
subclass pattern significantly (p0.001) increased the variance
explained in HDL cholesterol and log
e
triglyceride concentra-
tions to 14.1% for each. Body mass index alone explained
10.2% (p0.001) of the variance in HDL cholesterol and 5.9%
(p0.001) of the variance in log
e
triglyceride concentration.
The combination of body mass index and LDL subclass pattern
explained 16.3% of the variance in HDL cholesterol and 15.6%
of the variance in log
e
triglyceride concentration (p0.001 for
the additional variance explained by LDL subclass pattern in
both models).
Analysis of variance using waist or body mass index des-
ignated low and high based on a median split for the entire
study sample and LDL subclass distribution pattern as inde-
pendent variables was performed with lipid values as depen-
dent variables. Results from these analyses are shown in Table
4. No significant main effects were present for anthropometric
measures or LDL subclass pattern with respect to LDL choles-
terol concentration. Significant main effects were present for
anthropometric measures (waist and body mass index) and
LDL subclass pattern for both HDL cholesterol and triglycer-
ides. As was the case for the linear regression analysis, the
interaction terms for waist or body mass index with LDL
subclass pattern did not reach the 5% level of significance with
regard to HDL cholesterol or triglyceride concentrations (p
values all 0.14 or higher).
DISCUSSION
The results of the present investigation support the hypoth-
esis that postmenopausal women with LDL subclass pattern B
would have greater disturbances of the plasma lipid profile than
women with LDL pattern A. These results concur with those
for men published by Katzel and colleagues [23]. Both studies
showed that LDL subclass pattern B was associated with higher
levels of triglycerides and lower HDL cholesterol for a given
degree of adiposity. Differences were apparent even among
those with body mass index and waist circumference in the
non-obese range. These cross-sectional studies add support to
the hypothesis that the genetic factors which predispose to
expression of the small, dense LDL phenotype enhance the
deleterious effects of increased adiposity on triglyceride and
HDL cholesterol concentrations in men and women.
The lipid profile which characterizes LDL subclass pattern
B (elevated triglycerides, depressed HDL cholesterol and a
predominance of small, dense LDL particles) reflects an un-
derlying metabolic state which may influence responsiveness to
preventive therapies and prove useful for guiding treatment
selection. Katzel and colleagues [23] found that a 10 kg weight
Fig. 1. Results of regression analyses for the relationship between waist
circumference and log
e
triglycerides according to LDL subclass distri-
bution pattern.
Fig. 2. Results of regression analyses for the relationship between waist
circumference and HDL cholesterol according to LDL subclass distri-
bution pattern.
LDL Subclass Pattern B
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION 27
loss produced changes of 15% and 34% in triglyceride
levels among obese men with LDL subclass patterns A and B,
respectively (p0.01). However, the increase in HDL choles-
terol was smaller among men with LDL pattern B (16% vs.
10%, p0.05).
Dreon et al. [24] showed that men with LDL subclass
pattern B while consuming a reference diet high in fat (46% of
energy) had a more favorable plasma lipid response upon
switching to a low-fat diet (24% of energy) than did men with
pattern A during the reference diet phase. Men with pattern B
had larger reductions in LDL cholesterol and apolipoprotein B
and a trend toward a smaller decline in HDL cholesterol. The
same group of investigators showed that the LDL cholesterol
response to switching from a self-selected diet to one low in fat
and high in carbohydrate differed according to parental LDL
subclass pattern in premenopausal women [25]. Women with
two parents having LDL subclass pattern B showed the largest
LDL cholesterol change (36 mg/dL). Those with one pattern
B parent showed an intermediate response (9 mg/dL), while
the LDL response was minimal (2 mg/dL) in women whose
parents both had LDL subclass pattern A. In the Stanford
Coronary Risk Intervention Program, intensive risk factor mod-
ification (vs. usual care) retarded coronary artery disease pro-
gression among subjects with a predominance of dense LDL
(subclass pattern B) at baseline (0.008 vs. 0.054 mm/y,
p0.007). No benefit was observed among subjects with a
predominance of buoyant LDL upon entry (0.038 vs. 0.0039
mm/y) [26].
The present study shows a clear, additive influence of LDL
subclass pattern B on the dyslipidemia associated with in-
creased adiposity. Trends were also present in this sample
toward multiplicative interactions, i.e., greater worsening of the
lipid profile (HDL cholesterol and triglycerides) with increas-
ing waist circumference or body mass index among women
with LDL subclass pattern B, compared to those with pattern A.
The women studied were taking part in a clinical trial, and
the range of adiposity in the sample was restricted because a
trial exclusion criterion prevented enrollment of women with
body mass index 31.5 kg/m
2
. Restriction of the range of
adiposity would tend to reduce the power to detect non-paral-
lelism in the regression lines. Therefore, it is likely that the
failure to detect a significant multiplicative interaction is due to
Table 4. Serum Lipid Values according to Anthropometric Indicators and LDL Subclass Distribution Pattern
Independent Variables
Anthropometric Variable Main Effects
Low High
p Value
Waist or BMI
p Value
LDL Pattern
LDL Cholesterol (mg/dL)
MeanSEM
Waist 0.677 0.746
Pattern A 155.63.7 154.24.2
Pattern B 154.77.0 151.75.7
BMI 0.744 0.749
Pattern A 153.94.0 155.63.9
Pattern B 155.76.8 150.75.4
HDL Cholesterol (mg/dL)
MeanSEM
Waist 0.007 0.001
Pattern A 57.21.6 53.21.3
Pattern B 50.82.6 44.51.5
BMI 0.001 0.001
Pattern A 58.41.7 52.91.2
Pattern B 51.02.4 44.01.5
Triglycerides (mg/dL)
Geometric Mean (95% CI)
Waist 0.004 0.001
Pattern A 95.4 103.4
(88.2, 102.5) (95.6, 112.2)
Pattern B 115.9 147.8
(97.2, 138.2) (131.6, 165.7)
BMI 0.029 0.001
Pattern A 93.4 103.9
(86.5, 101.5) (96.0, 112.3)
Pattern B 124.3 141.7
(103.5, 148.4) (127.7, 157.6)
Abbreviations: BMIbody mass index; HDLhigh-density lipoprotein; LDLlow-density lipoprotein.
LDL Subclass Pattern B
28 VOL. 19, NO. 1
insufficient statistical power. Additional research will be
needed covering a wider range of adiposity to more fully
characterize the influence of LDL subclass pattern on the
dyslipidemia associated with increased adiposity.
A secondary objective of the current study was to compare
the utility of waist circumference and body mass index for
predicting adiposity-related alterations in the serum lipid pro-
file. Body mass index reflects total adiposity, whereas waist
circumference is a measure of both total and abdominal adi-
posity. These two measures were strongly correlated in our
sample (r0.77, p0.001), and both were significantly asso-
ciated with increased triglyceride and depressed HDL choles-
terol concentrations. Neither waist circumference nor body
mass index correlated significantly with the LDL cholesterol
level.
Our group has previously shown that the LDL cholesterol
concentration was directly related to measures of adiposity in a
group of younger (18 to 49 years) men, but that this relationship
was absent in men 50 and older [27]. Most women in the
present sample were 50 years of age or older. Thus, adiposity
does not appear to be a determinant of the LDL cholesterol
level among older persons of either gender. Based on these data
it might be anticipated that weight loss would not produce the
same degree of LDL cholesterol-lowering among older indi-
viduals that it does in young adults. Indeed, a meta-analysis of
trials investigating blood lipid responses to weight loss showed
that the mean LDL cholesterol response was larger for younger
subjects (25 mg/dL) than those middle-aged or older (8
mg/dL) [28].
CONCLUSIONS
In the current study, LDL subclass pattern B and anthropo-
metric indicators of adiposity (waist or body mass index) were
independent predictors of HDL cholesterol and triglyceride
levels in postmenopausal women. However, no significant re-
lationship was observed between LDL cholesterol and mea-
sures of adiposity. Non-significant trends were present toward
greater worsening of HDL cholesterol and triglyceride levels
with increasing adiposity among women with LDL subclass
pattern B, compared to those with pattern A. Body mass index
and waist circumference showed similar relationships to tri-
glyceride and HDL cholesterol concentrations, suggesting that
either measurement may be used for assessing the risk of
adiposity-related dyslipidemia in postmenopausal women.
REFERENCES
1. Austin MA, Breslow JL, Hennekens CH, Buring JE, Willett WC,
Krauss RM: Low-density lipoprotein subclass patterns and risk of
myocardial infarction. JAMA 260:19171921, 1988.
2. 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 276:875881, 1996.
3. Lamarche B, Tchernof A, Moorjani S, Cantin B, Dagenais GR,
Lupien PJ, Despres J-P: Small, dense low-density lipoprotein par-
ticles as a predictor of the risk of ischemic heart disease in men.
Prospective results from the Quebec Cardiovascular Study. Circu-
lation 95:6975, 1997.
4. Stampfer MJ, Krauss RM, Ma J, Blanche PJ, Holl LG, Sacks FM,
Hennekens CH: A prospective study of triglyceride level, low-
density lipoprotein particle diameter, and risk of myocardial in-
farction. JAMA 276:882888, 1996.
5. Reaven GM, Chen Y-D, Jeppesen J, Maheux P, Krauss RM:
Insulin resistance and hyperinsulinemia in individuals with small,
dense, low-density lipoprotein particles. J Clin Invest 92:141146,
1993.
6. Superko HR: What can we learn about dense low density lipopro-
tein and lipoprotein particles from clinical trials? Curr Opin Lipid
7:363368, 1996.
7. Halle M, Berg A, Keul J, Baumstark MW: Association between
serum fibrinogen concentrations and HDL and LDL subfraction
phenotypes in healthy men. Arterioscler Thromb Vasc Biol 16:
144148, 1996.
8. Krauss RM: Dense low density lipoproteins and coronary artery
disease. Am J Cardiol 75:53B57B, 1995.
9. Slyper AH: Low-density lipoprotein density and atherosclerosis.
Unraveling the connection. JAMA 272:305308, 1994.
10. Austin MA, Hokanson JE, Edwards KL: Hypertriglyceridemia as a
cardiovascular risk factor. JACC 81:7B12B, 1998.
11. Lamon-Fava S, Jimenez D, Christian JC, Fabsitz RR, Reed T,
Carmelli D, Castelli WP, Ordovas JM, Wilson PW, Schaefer EJ:
The NHLBI Twin Study: heritability of apolipoprotein A-I, B, and
low density lipoprotein subclasses and concordance of lipopro-
tein(a). Atherosclerosis 91:97106, 1991.
12. Haffner SM, Mykkanen L, Robbins D, Valdez R, Miettinen H,
Howard BV, Stern MP, Bowsher R: A preponderance of small
dense LDL is associated with specific insulin, proinsulin and the
components of the insulin resistance syndrome in nondiabetic
subjects. Diabetologia 138:13281336, 1995.
13. Selby JV, Austin MA, Newman B, Zhang D, Quesenberry CP,
Mayer EJ, Krauss RM: LDL subclass phenotypes and the insulin
resistance syndrome in women. Circulation 88:381387, 1993.
14. Tchernof A, Lamarche B, Prudhomme D, Nadeau A, Moorjani S,
Labrie F, Lupien PJ, Despres J-P: The dense LDL phenotype.
Association with plasma lipoprotein levels, visceral obesity, and
hyperinsulinemia in men. Diabetes Care 6:629637, 1996.
15. Katzel LI, Krauss RM, Goldberg AP: Relations of plasma TG and
HDL-C concentrations to body composition and plasma insulin
levels are altered in men with small LDL particles. Arterioscler
Thromb 14:11211128, 1994.
16. Austin MA, Newman B, Selby JV, Edwards K, Mayer EJ, Krauss
RM: Genetics of LDL subclass phenotypes in women twins. Con-
cordance, heritability, and comingling analysis. Arterioscler
Thromb 13:687695, 1993.
17. 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 36:2291
2302, 1995.
LDL Subclass Pattern B
JOURNAL OF THE AMERICAN COLLEGE OF NUTRITION 29
18. Kulkarni KR, Garber DW, Marcovina SM, Segrest JP: Quantifi-
cation of cholesterol in all lipoprotein classes by the VAP-II
method. J Lipid Res 35:159168, 1994.
19. DeLong DM, DeLong ER, Wood PD, Lippel K, Rifkind BM: A
comparison of methods for the estimation of plasma low- and very
low density lipoprotein cholesterol. The Lipid Research Clinics
Prevalence Study. JAMA 256:23722377, 1986.
20. Kriska AM, Caspersen CJ (eds): A collection of physical activity
questionnaires for health-related research. Med Sci Sports Exerc
29:S1S205, 1997.
21. Peters JR, Quiter ES, Brekke ML, Admire J, Brekke MJ, Mullis
RM, Hunninghake DB: The Eating Patterns Assessment Tool: a
simple instrument for assessing dietary fat and cholesterol intake.
J Amer Dietetic Assoc 94:10081013, 1994.
22. Kleinbaum DG, Kupper LL, Muller KE: Applied Regression
Analysis and Other Multivariable Methods, 2nd ed. Belmont, CA:
Wadsworth, 1988.
23. Katzel LI, Coon PJ, Rogus E, Krauss RM, Goldberg AP: Persis-
tence of low HDL-C levels after weight reduction in older men
with small LDL particles. Arterioscler Thromb Vasc Biol 15:299
305, 1995.
24. Dreon DM, Fernstrom HA, Miller B, Krauss RM: Low-density
lipoprotein subclass patterns and lipoprotein response to a reduced-
fat diet in men. FASEB J 8:121126, 1994.
25. Dreon DM, Fernstrom HA, Williams PT, Krauss RM: LDL sub-
class patterns and lipoprotein response to a low-fat, high-
carbohydrate diet in women. Arterioscler Thromb Vasc Biol 17:
707714, 1997.
26. Miller BD, Alderman EI, Haskell WL, Fair JM, Krauss RM:
Predominance of dense low-density particles predicts angiographic
benefit of therapy in the Stanford Coronary Risk Project. Circula-
tion 94:21462153, 1996.
27. Maki KC, Kritsch K, Foley S, Soneru I, Davidson MH: Age-
Dependence of the relationship between adiposity and serum low
density lipoprotein cholesterol in men. J Am Coll Nutr 16:578
583, 1997.
28. Datillo AM, Kris-Etherton PM: Effects of weight reduction on
blood lipids and lipoproteins, a meta-analysis. Am J Clin Nutr
56:320328, 1992.
Received June 1999; revision accepted November 1999.
LDL Subclass Pattern B
30 VOL. 19, NO. 1