Effects of Insulin Sensitivity, Body Composition, and Fitness on Lipoprotein Particle Sizes and Concentrations Determined by Nuclear Magnetic Resonance

Atherogenic lipoprotein profiles were associated with low levels of insulin sensitivity, cardiorespiratory fitness, and higher levels of adiposity, with truncal fat being the strongest predictor.

Abstract

Context:

Insulin resistance has been reported to be associated with development of atherogenic dyslipidemia. However, the confounding effects that obesity and low levels of cardiorespiratory fitness have on the relationship between insulin resistance and the development of atherogenic dyslipidemia remain to be adequately addressed.

Objective:

This study sought to examine the independent and combined effects of insulin sensitivity, body composition, and cardiorespiratory fitness on lipoprotein particle sizes and concentrations.

Methods:

Eight-four healthy, nondiabetic men (n = 43) and women (n = 41) were studied. The participants had a wide range of ages (18–30 and 65–80 yr), body composition (7.2–52.8% fat), and cardiorespiratory fitness (VO₂ peak, 13.5–66.2 ml/kg·min). Body composition, cardiorespiratory fitness, insulin sensitivity, and lipoprotein particle profiles were assessed using dual-energy x-ray absorptiometry, cardiopulmonary exercise testing, a hyperinsulinemic-euglycemic clamp, and nuclear magnetic resonance spectroscopy, respectively.

Results:

Low levels of insulin sensitivity and cardiorespiratory fitness and higher levels of adiposity were associated with the accumulation of small, dense, low-density lipoprotein particles; small high-density lipoprotein particles; triglycerides; and very low-density lipoprotein particles. Multivariate forward-stepwise regression revealed that higher levels of adiposity, in particular truncal fat, were the strongest predictor of the lipoprotein particle size and concentration data, followed by insulin sensitivity.

Conclusions:

As expected, the accumulation of atherogenic lipoprotein particles (e.g. small, dense, low-density lipoprotein particles and small, high-density lipoprotein particles) was associated with low levels of insulin sensitivity, cardiorespiratory fitness, and higher levels of adiposity. However, multivariate forward-stepwise regression revealed that triglycerides, followed by truncal fat mass, were the strongest predictors of the lipoprotein particle size and concentration data.

Dyslipidemia is a major determinant of atherosclerosis and is concomitantly associated with premature cardiovascular disease (CVD) and all-cause mortality. Insulin resistance is associated with high triglyceride and low high-density lipoprotein (HDL) cholesterol concentrations (1). However, diverse ranges of lipoprotein particle sizes exist among the major lipoprotein classes [e.g. HDL, low-density lipoprotein (LDL), and very low-density lipoprotein (VLDL)]. Emerging evidence indicates that lipoprotein particle size may play an important role in the manifestation of the atherogenic potential of a given lipoprotein particle (2, 3). Specifically, the accumulation of small LDL and HDL particles is associated with increased risk for atherosclerosis and premature CVD (4–6).

Using a hyperinsulinemic-euglycemic clamp and nuclear magnetic resonance (NMR) spectroscopy, Garvey et al. (7) demonstrated that insulin resistance is associated with an accumulation of these atherogenic lipoprotein particles. However, the confounding effects that obesity and low levels of cardiorespiratory fitness have on the relationship between insulin resistance and lipoprotein particles has not been fully addressed. Indeed, obesity and low cardiorespiratory fitness are both associated with insulin resistance (1, 8), as well as atherogenic dyslipidemia (8–10). Therefore, the purpose of the present investigation was to examine the independent and combined effects of insulin sensitivity, body composition, and cardiorespiratory fitness on lipoprotein particle sizes and concentrations determined using NMR spectroscopy.

Subjects and Methods

Healthy, nondiabetic men and women were recruited from the local community to participate in one of two protocols (11, 12) that were approved by the Mayo Clinic Institutional Review Board. The purpose, benefits, and risks of participation were fully explained before consent was obtained. Participants were excluded if they were on statins or antidiabetic medications. The participants had a wide range of ages (18–30 and 65–80 yr), body composition (7.2–52.8% fat), and cardiorespiratory fitness (VO₂ peak, 13.5–66.2 ml/kg·min). Body composition, cardiorespiratory fitness, whole-body insulin sensitivity, and lipoprotein particle profiles were assessed in the Mayo Clinic's Center for Translational Science Activities' Clinical Research Unit as described below.

Body composition was measured using dual-energy x-ray absorptiometry (Lunar DPX-L; Lunar Radiation, Madison, WI), and VO₂ peak was determined from indirect calorimetry during a graded bicycle test (11).

A hyperinsulinemic-euglycemic clamp was performed, infusing 1.5 mU/kg FFM/min insulin while maintaining similar plasma glucose levels [∼5.0 mmol/liter (90 mg/dl)] in every participant to assess whole-body insulin sensitivity. A mixture of amino acids (10% Travasol, Baxter Healthcare Corporation, Deerfield, IL) was infused to prevent insulin-induced hypoaminoacidemia. Arterialized venous blood was used to measure glucose levels every 10 min with a Beckman glucose analyzer (Beckman Coulter, Fullerton, CA). The glucose (40% solution) infusion rate (GIR) was adjusted to maintain euglycemia during the insulin infusion.

The lipoprotein particle size and concentrations were measured in EDTA plasma collected in the fasted state. The samples were analyzed by the Mayo Cardiovascular Medicine Laboratory using a 400-MHz NMR lipoprotein analyzer with size assignments made using the LipoProfile 2 software (LipoScience, Raleigh, NC), as previously described (13).

Subject demographics and metabolic parameters are presented as the median with the interquartile range by sex. Comparisons between females and males were conducted using nonparametric Wilcoxon rank sum tests. Nonparametric tests were conducted because the lipoprotein particle data were not normally distributed. Partial Spearman rank correlation analyses (adjusted for age group) were conducted to examine the association between insulin sensitivity (GIR), body composition (percentage fat and truncal fat mass), and fitness (VO₂ peak) with the lipoprotein particle size and concentration data. Multivariable, stepwise forward-selection (PROC REG, sle = 0.15) linear regression analyses were used to examine the association between the predetermined candidate predictors (independent variables), which included GIR, percentage body fat, truncal fat mass, VO₂ peak, age, and sex on the lipoprotein particle size and concentration data (dependent variables).

Results

Table 1 presents the participants' anthropometric, metabolic, and lipoprotein particle profiles by sex. As expected, the males were taller and heavier, while having lower percentage fat and higher VO₂ peak (all P < 0.05). In contrast, both males and females had similar levels of insulin sensitivity (GIR). Total cholesterol, non-HDL cholesterol, LDL cholesterol, LDL particle, and small VLDL particle concentrations were lower in males than in females (all P < 0.05). Table 2 presents the partial Spearman rank correlations adjusted for age group (young vs. elderly) by sex for the associations between insulin sensitivity (GIR), body composition (percentage fat and truncal fat mass), and cardiorespiratory fitness (VO₂ peak) with the lipoprotein particle size and concentration data.

Table 1.

Participant characteristics and NMR-derived lipoprotein particle profiles by sex

Variable label	Females (n = 41)	Males (n = 43)	P value	Diameter (nm)
Age (yr)	29 (23, 67)	29 (23, 67)	0.94
Height (cm)	163 (159, 171)	178 (172, 183)	<0.001
Weight (kg)	64.2 (55.2, 78.7)	78.0 (71.3, 86.5)	<0.001
BMI (kg/m2)	23.8 (20.8, 28.4)	24.5 (23.2, 27.7)	0.298
Fat (%)	35.6 (28.3, 43.6)	22.6 (15.5, 26.3)	<0.001
Truncal fat (kg)	9.4 (7.5, 15.1)	9.6 (5.4, 12.0)	0.090
VO2 peak (ml/kg·min)	24.2 (19.8, 31.4)	33.7 (27.5, 43.3)	<0.001
VO2 peak (ml/kg FFM·min)	42.3 (34.2, 50.3)	47.6 (38.2, 56.8)	0.076
GIR (μmol/kg FFM·min)	55.5 (48.8, 65.4)	54.9 (41.9, 64.2)	0.403
NMR cholesterol
NMR total cholesterol (mg/dl)	173 (152, 201)	151 (137, 172)	0.004
NMR non-HDL cholesterol (mg/dl)	129 (108, 149)	103 (94, 123)	0.003
LDL-cholesterol
Total LDL (mg/dl)	112 (86, 128)	90 (79, 105)	0.006
Total LDL particles (nmol/liter)	1099 (908, 1360)	965 (760, 1125)	0.024
IDL	25 (9, 76)	11 (0, 31)	0.012	23-27
Large LDL particles (nmol/liter)	520 (317, 699)	435 (350, 590)	0.161	21.2-23
Total small LDL particles (nmol/liter)	399 (242, 866)	521 (233, 715)	0.649	18-21.2
Medium-small LDL particles (nmol/liter)	94 (53, 180)	108 (52, 148)	0.943	19.8-21.2
Very small LDL particles (nmol/liter)	306 (199, 711)	413 (182, 589)	0.265	18-19.8
LDL mean particle size (nm)	21.6 (20.8, 22.0)	21.4 (20.8, 22.2)	0.968
HDL-cholesterol
Total HDL cholesterol (mg/dl)	47.0 (42.0, 56.0)	46.0 (41.0, 51.0)	0.325
Total HDL particles (μmol/liter)	31.8 (28.4, 34.7)	30.0 (27.5, 33.3)	0.323
Large HDL particles (μmol/liter)	7.1 (5.3, 8.8)	6.3 (4.0, 8.8)	0.304	8.8-13
Intermediate HDL particles (μmol/liter)	1.1 (0.2, 2.9)	2.2 (0.4, 5.2)	0.242	8.2-8.8
Small HDL particles (μmol/liter)	23.4 (18.2, 25.1)	20.8 (17.9, 25.3)	0.362	7.3-8.2
HDL mean particle size (nm)	9.0 (8.7, 9.3)	8.9 (8.6, 9.4)	0.862
Triglycerides
Triglycerides (mg/dl)	96.0 (80.0, 138.0)	86.0 (66.0, 121.0)	0.117
VLDL triglycerides (mg/dl)	59.0 (46.0, 99.0)	55.0 (38.0, 86.0)	0.330
Total VLDL particles (nmol/liter)	56.4 (43.4, 96.3)	44.3 (31.9, 73.5)	0.053
Large VLDL and chylomicron particles (nmol/liter)	1.7 (0.9, 3.0)	2.1 (0.9, 3.6)	0.819	>60
Medium VLDL particles (nmol/liter)	17.1 (9.9, 31.1)	12.2 (6.9, 22.4)	0.183	35-60
Small VLDL particles (nmol/liter)	37.3 (28.6, 53.3)	28.6 (20.4, 40.5)	0.041	27-35
VLDL mean particle size (nm)	49.0 (46.4, 54.5)	51.8 (47.0, 58.7)	0.203

Data are presented as median (interquartile range). Differences between sexes were tested using the Wilcoxon rank sum test. IDL, Intermediate-density lipoprotein.

Table 2.

Partial Spearman rank correlations adjusted for age

Variable label	Females (n = 41)				Males (n = 43)
	GIR	% Fat	Truncal fat	VO2 peak	GIR	% Fat	Truncal fat	VO2 peak
Total cholesterol (mg/dl)	0.201	0.019	−0.126	0.074	−0.118	0.364a	0.329a	−0.305a
Non-HDL cholesterol (mg/dl)	0.148	0.151	0.051	0.023	−0.297	0.562c	0.561c	−0.448c
LDL cholesterol
LDL cholesterol (mg/dl)	0.242	0.092	−0.051	0.099	−0.120	0.397b	0.352a	−0.313b
Total LDL particles (nmol/liter)	−0.063	0.428b	0.313a	−0.287	−0.361a	0.694c	0.710c	−0.454b
Large LDL particles (nmol/liter)	0.406a	−0.316a	−0.349a	0.430b	0.344a	−0.184	−0.273	0.009
Total small LDL particles (nmol/liter)	−0.371a	0.570c	0.505c	−0.525c	−0.396a	0.617c	0.662c	−0.367a
Medium-small LDL particles (nmol/liter)	−0.394a	0.545c	0.482b	−0.516c	−0.391a	0.567c	0.615c	−0.340a
Very small LDL particles (nmol/liter)	−0.354a	0.566c	0502c	−0.516c	−0.403b	0.629c	0.681c	−0.370a
LDL mean particle size (nm)	0.386	−0.514c	−0.476b	0.522c	0.377a	−0.518c	−0.577c	0.292
HDL cholesterol
HDL cholesterol (mg/dl)	0.210	−0.304	−0.399a	0.255	0.508c	−0.464c	−0.581c	0.309a
Total HDL particles (μmol/liter)	−0.017	−0.071	−0.154	0.094	−0.115	0.303a	0.286	−0.254
Large HDL particles (μmol/liter)	0.266	−0.330a	−0.405b	0.280	0.489c	−0.522c	−0.647c	0.351a
Medium HDL particles (μmol/liter)	−0.262	0.024	−0.014	−0.176	−0.322a	0.153	0.217	−0.218
Small HDL particles (μ mol/liter)	−0.082	0.171	0.198	0.054	−0.245	0.472b	0.505c	−0.275
HDL mean particle size (nm)	0.386a	−0.464b	−0.465b	0.400a	0.592c	−0.620c	−0.755c	0.449b
Triglycerides
Triglycerides (mg/dl)	−0.344a	0.374a	0.292	−0.369a	−0.598c	0.623c	0.745c	−0.521c
VLDL triglycerides (mg/dl)	−0.411b	0.385a	0.293	−0.420b	−0.583c	0.587c	0.712c	0.491c
Total VLDL particles (nmol/liter)	−0.246	0.233	0.153	−0.240	−0.482b	0.547c	0.616c	−0.442a
Large VLDL and chylomicron particles (nmol/liter)	−0.425b	0.342a	0.278	−0.367a	−0.524c	0.369a	0.567c	−0.370a
Medium VLDL particles (nmol/liter)	−0.346a	0.247	0.172	−0.333a	−0.448b	0.485c	0.539c	−0.288
Small VLDL particles (nmol/liter)	−0.051	0.123	0.094	−0.064	−0.291	0.456b	0.468c	−0.444b
VLDL mean particle size (nm)	−0.278	0.244	0.162	−0.312	−0.181	−0.019	0.128	−0.118

VO₂ peak, Peak pulmonary oxygen uptake (ml/kg/min). GIR units, μmol/kg FFM/min. % Fat and truncal fat mass (in kilograms) were determined by dual-energy x-ray absorptiometry. Bold indicates P < 0.05.

^aP < 0.05;

^bP < 0.01;

^cP < 0.001.

Higher GIRs were associated with lower concentrations of medium-small LDL particles, very small LDL particles, and total small LDL particles, with concomitantly higher concentrations of large LDL particles and larger mean LDL particle size independent of sex. Moreover, higher GIR was also associated with lower concentrations of triglycerides, VLDL triglycerides, VLDL particles, large VLDL, and chylomicron particles, and medium VLDL particles also independent of sex. In contrast, there was a difference between males and females with respect to the positive association between GIR and large HDL particles observed in men, but not in women.

A higher percentage of body fat and truncal fat mass were associated with higher concentrations of total cholesterol and non-HDL cholesterol in men, but not in women. A higher percentage of body fat and truncal fat mass were associated with higher concentrations of total LDL particles, medium-small LDL particles, very small LDL particles, and total small LDL particles, and smaller mean LDL particle size independent of sex. In addition, higher percentage of body fat and truncal fat mass were associated with lower concentrations of total HDL cholesterol, large HDL particles, and smaller mean HDL particle size independent of sex. Moreover, higher percentage of body fat and truncal fat mass were also associated with higher concentrations of triglycerides, VLDL triglycerides, and large VLDL and chylomicron particles independent of sex.

A higher VO₂ peak was associated with lower concentrations of total cholesterol, non-HDL cholesterol, LDL cholesterol, and total LDL particles in men, but not in women. A higher VO₂ peak was associated with lower concentrations of medium-small LDL particles and very small LDL particles independent of sex. A higher VO₂ peak was associated with higher concentrations of HDL cholesterol and large HDL particles in males, but not in females. Moreover, higher VO₂ peak was also associated with lower concentrations of triglycerides, VLDL triglycerides, VLDL particles, and large VLDL and chylomicron particles, and small VLDL particles independent of sex.

Supplemental Table 1 (published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org) presents the results of the multivariate forward-stepwise model. The results indicate that body composition (in particular, truncal fat mass) was the strongest predictor of the lipoprotein particle size and concentration data. Indeed, truncal fat mass accounted for approximately 22–30% of the variability in medium to small LDL particles. The GIR was also a significant predictor of the lipoprotein particle size and concentration data. However, the GIR typically accounted for less than 10% of the variability in the lipoprotein particle data.

Discussion

The main findings of the current study are that, in nondiabetic adults: 1) lower levels of insulin sensitivity and cardiorespiratory fitness and higher levels of adiposity were associated with higher concentrations of atherogenic lipoprotein particles (e.g. very small LDL particles and small HDL particles) and smaller mean LDL and HDL particle sizes; and 2) triglycerides, truncal fat mass, and insulin sensitivity (GIR) were the strongest predictors of the lipoprotein particle size and concentrations.

Insulin resistance is frequently associated with the development of atherogenic dyslipidemia, characterized by high triglyceride and low HDL-cholesterol concentrations (1). Moreover, insulin resistance is also associated with an accumulation of both small, dense LDL and small HDL particles (7, 14, 15). Using a hyperinsulinemic-euglycemic clamp and NMR spectroscopy, the present investigation indicates that low levels of insulin sensitivity were associated with an accumulation of small LDL particles and smaller mean LDL and HDL particle sizes. In addition, low levels of insulin sensitivity were also associated with an accumulation of VLDL triglycerides. Although the present data indicate that insulin resistance is associated with an accumulation of atherogenic lipoprotein particles, it remains unclear whether this association is independent of concomitant changes in body composition and/or cardiorespiratory fitness.

Insulin resistance is often associated with increased adiposity, which is an independent risk factor for atherogenic dyslipidemia (10, 16). Indeed, obesity, particularly abdominal obesity, is associated with an accumulation of small, dense LDL particles (16) and an accumulation of small HDL particles (10). In agreement with these previous results, the present data indicated that increased levels of whole-body and abdominal adiposity were also associated with an accumulation of small LDL particles, small HDL particles, and smaller mean LDL and HDL particle sizes. Moreover, increased levels of adiposity were also associated with adverse triglyceride and VLDL particle profiles, particularly in males.

The current results demonstrated that lower levels of cardiorespiratory fitness were associated with an accumulation of small LDL particles, and concomitant elevations in triglycerides and VLDL particles. Low cardiorespiratory fitness is also associated with insulin resistance (9), abdominal adiposity (8, 9, 17), and atherogenic dyslipidemia (8, 9, 17). Whether low cardiorespiratory fitness independently contributes to the development of atherogenic dyslipidemia remains unclear. Low cardiorespiratory fitness has recently been reported to be associated with several features of the metabolic syndrome, including both abdominal adiposity and atherogenic dyslipidemia (9). Moreover, increased levels of cardiorespiratory fitness have been reported to attenuate the risk of the metabolic syndrome (17). Recent studies have demonstrated that those who engage in regular endurance exercise and maintain higher VO₂ peak have favorable cardioprotective HDL and LDL particles (12).

Multivariate forward-stepwise regression revealed that body composition, in particular, truncal fat mass, was the strongest predictor of the lipoprotein particle size and concentration data (Supplemental Table 1). Insulin sensitivity was selected as the second most significant predictor of the lipoprotein particle size and concentration data. In contrast, cardiorespiratory fitness was not selected as a significant predictor of the lipoprotein particle size and concentration data. However, it has been shown that cardiorespiratory fitness is an important determinant of abdominal (visceral) fat (18). Therefore, it remains to be determined whether exercise and related elevations in cardiorespiratory fitness act to improve lipoprotein particles via their effect on visceral fat. However, when triglycerides were included as a candidate predictor in the multivariate forward-stepwise regression model, they emerged as the strongest predictor of the LDL and HDL particle size and concentration data (Supplemental Table 2).

Clinically, there is growing evidence that lipoprotein particle size plays a role in mediating CVD and metabolic risk (19). In particular, recent evidence suggests that it is not only the quantity of cholesterol levels but also the quality that counts in mediating the atherogenic potential of LDL and HDL particles (19). Indeed, concentrations of large LDL and HDL particles have been reported to be elevated in individuals with exceptional longevity and were inversely related to other CVD risk factors (20).

In summary, low levels of insulin sensitivity and cardiorespiratory fitness and higher levels of adiposity were associated with the presence of atherogenic dyslipidemia, marked by the accumulation of small, dense LDL particles, small HDL particles, triglycerides, and VLDL particles. However, multivariate forward-stepwise regression revealed that higher levels of triglycerides, adiposity (in particular, abdominal adiposity), and lower levels of insulin sensitivity were the strongest predictors of the lipoprotein particle size and concentration data.