Abstract
We examined the association between serum lipoprotein subclasses and the three measures of arterial stiffness i.e. (i) carotid-femoral pulse wave velocity (cfPWV) which is a gold standard measure of central arterial stiffness, (ii) brachial-ankle PWV (baPWV) which is emerging as a combined measure of central and peripheral arterial stiffness, and (iii) femoral-ankle PWV (faPWV) which is a measure of peripheral arterial stiffness. Among a population-based sample of 701 apparently healthy Caucasian, Japanese American and Korean men aged 40–49 years, concentrations of lipoprotein particles were assessed by nuclear magnetic resonance (NMR) spectroscopy, and PWV was assessed with an automated waveform analyzer (VP2000, Omron, Japan). Multiple linear regressions were performed to analyze the association between each NMR lipoprotein subclasses and PWV measures, after adjusting for cardiovascular risk factors and other confounders. A cut-off of p<0.01 was used for determining significance. All PWV measures had significant correlations with total and small low-density lipoprotein particle number (LDL-P) (all p<0.0001) but not LDL-cholesterol (LDL-C) (all p>0.1), independent of race and age. In multivariate regression analysis, no NMR lipoprotein subclass was significantly associated with cfPWV (all p>0.01). However, most NMR lipoprotein subclasses had significant associations with both baPWV and faPWV (p<0.01). In this study of healthy middle-aged men, as compared to cfPWV, both baPWV and faPWV had stronger associations with particle numbers of lipoprotein subclasses. Our results may suggest that both baPWV and faPWV are related to arterial stiffness and atherosclerosis, whereas cfPWV may represent arterial stiffness alone.
Keywords: lipoproteins, lipoprotein fractions, pulse wave velocity, atherosclerosis
Introduction
Carotid-femoral pulse-wave velocity (cfPWV) is established as a gold standard measure of central arterial stiffness, and is known to increase with age and BP.1 cfPWV is an independent predictor of CVD events as well as all-cause mortality.2 The European Society of Hypertension has recommended cfPWV as a marker for subclinical organ damage among patients with hypertension, and in identifying patients with high CVD risk.3 Brachial-ankle pulse wave velocity (baPWV), which is likely to be a measure of combined central and peripheral arterial stiffness, correlates well with cfPWV,4 and is increasingly being used clinically as a measure of cardiovascular risk.5 Both cfPWV and baPWV predict CVD events and all-cause mortality in the general population as well as among CVD patients, 2,5 and are shown to be markers of therapeutic progress.6 In contrast, femoral-ankle pulse wave velocity (faPWV), as the name implies, is the velocity at which the pulse wave travels from the femoral artery to the posterior tibial artery at the ankle. Thus, faPWV is likely to be an indicator of arterial stiffness in the peripheral arterial segment.
Nuclear magnetic resonance (NMR) spectroscopy has emerged as a technique for quantifying the size and number of particles within subclasses of serum lipoproteins. While serum total cholesterol and low-density lipoprotein cholesterol (LDL-C) are traditional measures of atherogenic serum lipoproteins,7 LDL-C as compared to NMR-derived low-density lipoprotein particle number (LDL-P) may be a less accurate determinant of lipemic atherosclerotic risk.8 For example, in a multi-ethnic cohort of 6,814 US adults, among participants with discordance between LDL-C and LDL-P, LDL-P had a stronger association with atherosclerosis.8
Although many studies have not found age-, BP-independent association between central arterial stiffness (cfPWV) and serum lipids such as serum total cholesterol, LDL-C or high density lipoprotein cholesterol (HDL-C),9 to our knowledge, no previous population-based study has examined the association between NMR lipoprotein subclasses and different measures of arterial stiffness. We examined the association of particle numbers of NMR lipoprotein subclasses with cfPWV, baPWV and faPWV in the Electron-beam computed tomography, Risk factor Assessment among Japanese and U.S. Men in the Post-World War II birth cohort (ERA JUMP Study), an international study of subclinical atherosclerosis among middle-aged men.
Materials and Methods
Participants
During 2002–2006, a population-based sample of 1,228 men aged 40–49 years, with no clinical CVD or type 1 diabetes, was obtained from 4 centers: 310 whites from Allegheny County, Pennsylvania; 303 Japanese Americans (JA) from Honolulu, Hawaii; 313 Japanese from Kusatsu, Shiga, Japan; and 302 Koreans from Ansan, Gyeonggi-do, South Korea as previously described.10 Japanese in Japan were excluded from this analysis as cfPWV was not examined. Of the original sample, we further excluded those with diabetes (n=82), taking lipid-lowering medications (n=112) or anti-hypertensive medications (n=112). Diabetes was defined as individuals with fasting glucose ≥126 mg/dl or use of medications for diabetes. The final sample consisted of 701 men (248 whites, 191 JA, and 262 Koreans). Written informed consent was obtained from all participants. The study was approved by the Institutional Review Boards of the following institutions: the University of Pittsburgh, Pittsburgh, Pennsylvania; the Kuakini Medical Center, Honolulu, Hawaii; Shiga University of Medical Science, Otsu, Japan; and Korea University, Seoul, South Korea.
All participants underwent a physical examination, completed a lifestyle questionnaire and a laboratory assessment as described previously.10–12 Body weight and height were measured while the participant was wearing light clothing without shoes. Body-mass index (BMI) was calculated as weight in kilograms divided by the square of the height in meters. An Alcohol drinker was defined as intake ≥ twice/week. Smoking was determined by pack years of cigarette smoking. Blood pressure and heart rate were measured after the participant emptied his bladder and sat quietly for 5 min. Blood pressure was measured twice on both arms with an automated sphygmomanometer (BP-8800, Colin Medical Technology, Komaki, Japan). The average of the two measurements was used. Hypertension was defined as SBP ≥ 140 mmHg and/or DBP ≥ 90 mmHg. Venipuncture was performed early in the clinic visit after a 12-h fast. Blood samples were stored at −80°C and shipped on dry ice to the University of Pittsburgh. Serum lipids were determined using CDC-standardized methods.10 Serum glucose was determined using hexokinase–glucose-6-phosphate-dehydrogenase enzymatic assay. Data collection was standardized across all centers.
Nuclear magnetic resonance lipoprotein measurements
Serum samples obtained from study participants were shipped from the University of Pittsburgh to LipoScience (LipoScience Inc., Raleigh, NC) for NMR spectroscopy. NMR spectroscopy using LipoProfile-3 algorithm was performed to quantify the particle numbers of very low-density lipoprotein (VLDL-P), LDL-P, and high-density lipoprotein (HDL-P).13, 14 Particle concentrations were determined for 3 VLDL subclasses (large, >60 nm; medium, 35–60 nm; and small, 29–35 nm), 3 LDL subclasses (intermediate-density lipoprotein (IDL), 23–29 nm; large, 20.5–23 nm; small, 18–20.5 nm), and 3 HDL subclasses (large, 9.4–14 nm; medium, 8.2–9.4 nm; and small, 7.3–8.2 nm).
Measurement of PWV
At the start of the study, staff from the University of Pittsburgh’s Ultrasound Research Laboratory visited the Honolulu site to train the sonographers in Honolulu and from South Korea for PWV measurements. PWV measurements were automatically generated using a noninvasive and automated waveform analyzer (VP2000, Omron, Japan).4 This device recorded the carotid and femoral pulse waveform using multi-array tonometer, and provided automated measures of cfPWV, faPWV, and baPWV. Following 10 min of rest in a supine position, occlusion and monitoring cuffs were placed around both arms and both ankles of the participant. The arm cuffs were placed on skin or over light clothing, and the ankle cuffs were directly placed over the skin. ECG electrodes were placed on both wrists and a phonocardiogram i.e. a microphone for detecting heart sounds was placed on the left edge of the sternum. For cfPWV path length, the distance between the carotid and femoral sites was measured over the surface of the body with a tape measure. The path lengths for baPWV and faPWV were calculated using height-based formulae.15 Sonographers palpated the left femoral artery and the left carotid artery, and placed the handheld tonometer over these two pulse areas to obtain femoral and carotid pulse waveforms simultaneously. A foot-pedal was used to start the recording. Data was collected two times for each participant and the observed values were averaged. cfPWV was calculated by time-phase analysis using volume waveforms of the respective arteries (carotid and femoral arteries). PWV was calculated as distance between arterial sites divided by the time between the feet of the respective waveforms. Intra-class correlations (ICC) for re-examination were 0.76 (cfPWV), 0.97 (baPWV), and 0.96 (faPWV) within technicians. Similarly, ICCs of 0.73 (cfPWV), 0.87(faPWV), and 0.91 (baPWV) were achieved between technicians.
Statistical analysis
We first calculated overall and race-specific mean (standard deviation) or median (interquartile range) values of the study population, as appropriate. Frequencies and proportions were calculated for categorical variables. Analysis of variance and chi-square tests were used to examine significant differences between individual racial groups. Two-tailed cut-off p-value for significance was kept at 0.05 for these analyses.
We then used Spearman rank statistics to test the age-, race-adjusted correlation between each of individual lipoprotein subclasses and the different measures of PWV i.e. (i) cfPWV, (ii) baPWV, and (iii) faPWV. To determine its statistical significance, we used a more conservative two-tailed p-value of 0.01 to compensate for multiple comparisons. Multiple linear regressions were then used to calculate standardized regression coefficients and their significance, keeping individual NMR lipoprotein subclasses and serum lipids as predictor variables, and each PWV measure as an outcome variable. For the multiple linear regressions, other covariates adjusted for in the model were age, BMI, race, glucose, SBP, heart rate, pack years of smoking, and frequency of alcohol intake. Similar to previous analysis, we used a two-tailed p-value of 0.01 as cut-off for significance.
To test the presence of effect modification of lipoprotein subclasses by race, we tested for statistical interaction between race and predictor variable i.e. each of the lipoprotein subclasses in a regression model for the outcome variable i.e. cfPWV, baPWV and faPWV. Regression models testing for interactions were unadjusted for other covariates. We used the p-value of <0.05 for determining significant interaction in the regression model. All data analysis was performed using SAS/STAT® software, Version 9.3 of the SAS System, Cary, NC, USA.
Results
Table 1 presents descriptive characteristics of the study population overall, and by race. About 7.7% of the study population had SBP ≥140 mmHg; 45% were ever-smokers and the median value of pack-years of smoking among ever-smokers was 14 (interquartile range: 7, 24.5). More than two-fifths of the study participants were alcohol drinkers. Koreans, as compared to whites and JA, had significantly less BMI, SBP, total cholesterol, HDL-C and LDL-C, although they had significantly higher pack-years of smoking.
Table 1.
Descriptive characteristics of the study population1
| All subjects (n=701) | Whites (n=248) | Japanese Americans (n=191) | Koreans (n=262) | Difference between racial groups2 | |
|---|---|---|---|---|---|
|
| |||||
| Age | 45.1 (2.9) | 44.8 (2.8) | 46.0 (2.9) | 44.7 (2.8) | ¶¥ |
| BMI (kg/m2) | 26.2 (3.8) | 27.5 (4.3) | 26.7 (3.6) | 24.5 (2.6) | ഠ|
| Systolic blood pressure (mmHg) | 122.2 (12.0) | 122.2 (11.3) | 125.8 (12.1) | 119.5 (12) | ¶‡¥ |
| Diastolic blood pressure (mmHg) | 74.4 (9.2) | 72.6 (8.7) | 76.5 (8.9) | 74.6 (9.7) | ¶‡ |
| Hypertension (%) | 9.4 | 7.7 | 15.2 | 6.9 | ¶¥ |
| Heart rate | 65.4 (9.0) | 64.1 (9.3) | 66.4 (8.5) | 65.8 (9.1) | ¶ |
| Glucose (mg/dL) | 101.5 (11.7) | 99 (7.8) | 105.5 (7.8) | 100.9 (15.7) | ¶¥ |
| Smoking (pack-years) | 7.5 (12.0) | 3.7 (8.5) | 4.0 (8.1) | 13.9 (14.4) | ഠ|
| Alcohol drinker3 | 42.1 | 45.6 | 37.5 | 42.2 | none |
| Total cholesterol (mg/dL) | 205.4 (36.7) | 215.3 (37.4) | 211.9 (35.6) | 191.6 (32.6) | ഠ|
| LDL cholesterol (mg/dL) | 127.0 (33.6) | 137.6 (33.2) | 129.1 (31.9) | 115.4 (31.5) | ¶‡¥ |
| HDL cholesterol (mg/dL) | 48.3 (12.6) | 48.7 (13.1) | 51.0 (12.5) | 45.9 (11.9) | ഠ|
| Triglycerides (mg/dL) | 153.5 (96.5) | 145 (80.7) | 163.6 (111.9) | 154.1 (97.6) | none |
| NMR Lipoproteins | |||||
| Total VLDL-P (nmol/L) | 81.2 (46.1) | 82.8 (43.7) | 95.4 (51.6) | 69.4 (40.9) | ¶‡¥ |
| Large VLDL-P (nmol/L) | 5.1 (7.4) | 5.5 (7.4) | 6.3 (8.7) | 3.9 (6.0) | ¥ |
| Medium VLDL-P (nmol/L) | 35.1 (30.8) | 36.9 (29.3) | 42.1 (34.9) | 28.3 (27.6) | ഠ|
| Small VLDL-P (nmol/L) | 41.1 (25.3) | 40.5 (24.1) | 47.1 (26.8) | 37.3 (24.5) | ¶¥ |
| Total LDL-P (nmol/L) | 1309.4 (347.0) | 1474.4 (338.39) | 1290.8 (334.8) | 1166.9 (294.0) | ¶‡¥ |
| Intermediate DL-P (nmol/L) | 143.5 (96.5) | 131.6 (99.7) | 139.0 (100.0) | 158.0 (89.2) | ‡ |
| Large LDL-P (nmol/L) | 546.8 (279.9) | 665.0 (308.6) | 455.7 (245.5) | 501.3 (233.8) | ¶‡ |
| Small LDL-P (nmol/L) | 619.1 (339.4) | 677.7 (350.9) | 696.2 (338.0) | 507.5 (298.5) | ഠ|
| Total HDL-P (μmol/L) | 31.8 (6.4) | 31.5 (5.8) | 36.4 (5.7) | 28.7 (5.4) | ¶‡¥ |
| Large HDL-P (μmol/L) | 3.8 (3.0) | 3.3 (2.7) | 4.6 (3.1) | 3.8 (2.9) | ¶¥ |
| Medium HDL-P (μmol/L) | 7.8 (4.2) | 7.7 (4.0) | 10.2 (4.5) | 6.0 (3.1) | ¶‡# |
| Small HDL-P (μmol/L) | 20.2 (4.5) | 20.6 (4.4) | 21.5 (4.2) | 18.9 (4.3) | ‡¥ |
Values are expressed as mean (SD)
Participants taking antihypertensive (n=140), lipid-lowering (n=132) and hypoglycemic medications (n=33) were removed from the analysis (Total excluded = 305)
¶ denotes significant difference between whites and JA, ‡ denotes significant difference between whites and Koreans, and ¥ denotes significant difference between JA and Koreans
Alcohol drinker was defined as alcohol intake ≥ twice/week
BMI = body-mass index, LDL = low-density lipoprotein, HDL = high-density lipoprotein, VLDL = very low-density lipoprotein, NMR = nuclear magnetic resonance
Table 2 presents the three measures of PWVs in the three racial groups. Koreans had lower PWV than JA among all the three PWV measures, while they had lower cfPWV and faPWV than Whites. JA had higher PWVs than the other racial groups for all three measures.
Table 2.
Mean pulse wave velocity values in the study populations
| cfPWV (SD), cm/s | baPWV (SD), cm/s | faPWV (SD), cm/s | |
|---|---|---|---|
|
| |||
| Whites | 889.6 (268.3) | 1310.7 (150.4) | 987.5 (160.4) |
| Japanese Americans | 882.7 (189.6) | 1413.1 (179.0) | 1062.5 (111.7) |
| Koreans | 802.9 (128.8) | 1286.4 (137.2) | 1026.1 (98.0) |
| Total | 854.4 (205.2) | 1329.8 (162.6) | 1023.1 (129.0) |
|
| |||
| Difference between racial groups1 | ‡¥ | ¶¥ | ¶‡¥ |
¶ denotes significant difference between whites and JA, ‡ denotes significant difference between whites and Koreans, and ¥ denotes significant difference between JA and Koreans
cfPWV = carotid-femoral pulse wave velocity, baPWV = brachial-ankle pulse wave velocity, faPWV = femoral-ankle pulse wave velocity
Table 3 presents Spearman partial correlations of each subclass of lipoproteins with the three measures of PWVs, adjusted for age and race. cfPWV strongly correlated with both known atherogenic lipoprotein subclasses i.e. small LDL-P and total LDL-P (both p<0.001) even after adjusting for age and race, and also had a significant inverse correlation with large HDL-P (p=0.0007). In slight contrast, baPWV significantly correlated with most lipoprotein subclasses, while having significant inverse correlations with large LDL-P (p<0.0001). Similarly, faPWV had significant positive correlations with many lipoprotein subclasses, and a significant inverse correlation with large LDL-P (p=0.0002). Similar results were obtained while using Pearson’s correlation coefficient for testing correlation between NMR lipoproteins and PWV (data not shown).
Table 3.
Spearman Rank Correlations1 between measures of serum lipoproteins and pulse wave velocity in the study population
| cfPWV | baPWV | faPWV | ||||
|---|---|---|---|---|---|---|
|
| ||||||
| Spearman ρ | p-value | Spearman ρ | p-value | Spearman ρ | p-value | |
| Total VLDL-P | 0.1133 | 0.0034 | 0.2218 | <0.0001 | 0.1840 | <0.0001 |
| Large VLDL-P | 0.2351 | <0.0001 | 0.3326 | <0.0001 | 0.2898 | <0.0001 |
| Medium VLDL-P | 0.1517 | <0.0001 | 0.2555 | <0.0001 | 0.2128 | <0.0001 |
| Small VLDL-P | −0.0318 | 0.4120 | 0.0127 | 0.7425 | 0.0019 | 0.9602 |
| Triglycerides | 0.2176 | <0.0001 | 0.2763 | <0.0001 | 0.2325 | <0.0001 |
| VLDL Size | 0.2038 | <0.0001 | 0.2674 | <0.0001 | 0.2428 | <0.0001 |
| Intermediate DL-P | 0.0928 | 0.0164 | 0.0978 | 0.0112 | 0.0991 | 0.0105 |
| Total LDL-P | 0.1827 | <0.0001 | 0.1651 | <0.0001 | 0.1102 | 0.0044 |
| Large LDL-P | −0.0569 | 0.1415 | −0.1836 | <0.0001 | −0.1444 | 0.0002 |
| Small LDL-P | 0.1961 | <0.0001 | 0.2981 | <0.0001 | 0.1963 | <0.0001 |
| LDL-C | 0.0576 | 0.1370 | 0.0314 | 0.4170 | −0.0016 | 0.9665 |
| LDL Size | −0.1271 | 0.0010 | −0.2794 | <0.001 | −0.1951 | <0.0001 |
| Total HDL-P | 0.0701 | 0.0702 | 0.2552 | <0.0001 | 0.2186 | <0.0001 |
| Large HDL-P | −0.1302 | 0.0007 | −0.0746 | 0.0536 | −0.0187 | 0.6305 |
| Medium HDL-P | 0.0407 | 0.2938 | 0.1556 | <0.0001 | 0.0920 | 0.0176 |
| Small HDL-P | 0.1387 | 0.0003 | 0.2416 | <0.0001 | 0.2162 | <0.0001 |
| HDL-C | −0.0614 | 0.1125 | 0.0373 | 0.3345 | 0.0387 | 0.3178 |
| HDL Size | −0.1089 | 0.0048 | −0.0513 | 0.1843 | 0.0096 | 0.8038 |
Partial correlation after adjusting for age and race
Spearman ρ = Spearman rank correlation coefficient. Cut-off p-value of p<0.01 was used for determining significance.
Unit for VLDL-P and LDL-P was nmol/L, unit for HDL-P was μmol/L. Unit for pulse wave velocity was cm/s
cfPWV = carotid-femoral pulse wave velocity, baPWV = brachial-ankle pulse wave velocity, faPWV = femoral-ankle pulse wave velocity
VLDL-P = very low-density lipoprotein particle number, VLDL-C = very low-density lipoprotein cholesterol, LDL-P = low-density lipoprotein particle number, LDL-C = low-density lipoprotein cholesterol, HDL-P = high-density lipoprotein particle number, HDL-C = high-density lipoprotein cholesterol
Table 4 presents multivariable-adjusted linear regression analysis of PWV measures with each of individual lipoprotein subclasses as main predictors. Standardized regression coefficients are shown which represent SD change in the PWV measure with every 1 SD change in the NMR lipoprotein subclass. When we adjusted for risk factors of CVD such as BMI, SBP, glucose, smoking and alcohol drinking status, NMR lipoprotein subclasses had significant associations with baPWV and faPWV. In contrast, the significant associations between atherogenic lipoproteins and cfPWV, as seen in the partial correlation analysis, became insignificant after multivariate adjustment. None of the lipoprotein subclasses was significantly associated with cfPWV after multivariate adjustment (all p>0.01). In contrast, baPWV had a significant positive association with medium HDL-P (p=0.0068), small HDL-P (p<0.0001), total HDL-P (p<0.0001), small LDL-P (p<0.0001), large VLDL-P (p<0.0001), medium VLDL-P (p<0.0001), total VLDL-P (p<0.0001), and VLDL size (p<0.0001). In addition, baPWV was significantly inversely associated with large LDL-P (p<0.0001) and LDL size (p<0.0001). Similar to baPWV, faPWV had significant associations with lipoprotein subclasses, except that it was not significantly associated with medium HDL-P.
Table 4.
Association between NMR subclasses and carotid-femoral, brachial-ankle and femoral-ankle pulse wave velocities in the study population
| Standardized coefficients (β) | ||||||
|---|---|---|---|---|---|---|
|
| ||||||
| cfPWV1 | p-value | baPWV1 | p-value | faPWV1 | p-value | |
| Large VLDL-P | 0.0608 | 0.1113 | 0.1825 | <0.0001 | 0.1857 | <0.0001 |
| Medium VLDL-P | 0.0645 | 0.0850 | 0.1914 | <0.0001 | 0.1561 | <0.0001 |
| Small VLDL-P | −0.0339 | 0.3420 | 0.0152 | 0.6463 | 0.0310 | 0.3904 |
| Total VLDL-P | 0.0308 | 0.4069 | 0.1600 | <0.0001 | 0.1465 | <0.0001 |
| Triglycerides2 | 0.0687 | 0.0768 | 0.1650 | <0.0001 | 0.1541 | <0.0001 |
| VLDL Size | 0.0440 | 0.2543 | 0.1465 | <0.0001 | 0.1608 | <0.0001 |
| Intermediate DL-P | −0.0068 | 0.8523 | 0.0479 | 0.1573 | 0.0622 | 0.0910 |
| Large LDL-P | 0.0104 | 0.7826 | −0.1428 | <0.0001 | −0.0871 | 0.0218 |
| Small LDL-P | 0.0625 | 0.1125 | 0.1867 | <0.0001 | 0.1296 | 0.0011 |
| Total LDL-P | 0.0700 | 0.0798 | 0.0736 | 0.0473 | 0.0715 | 0.0771 |
| LDL-C | 0.0152 | 0.6824 | 0.0177 | 0.6091 | −0.0014 | 0.9711 |
| LDL Size | −0.0132 | 0.7297 | −0.1911 | <0.0001 | −0.1426 | 0.0002 |
| Large HDL-P | −0.0210 | 0.5966 | −0.0357 | 0.3339 | −0.0387 | 0.3364 |
| Medium HDL-P | −0.0028 | 0.9420 | 0.0954 | 0.0068 | 0.0402 | 0.2946 |
| Small HDL-P | 0.0545 | 0.1481 | 0.1584 | <0.0001 | 0.1742 | <0.0001 |
| Total HDL-P | 0.0313 | 0.4316 | 0.1759 | <0.0001 | 0.1477 | 0.0002 |
| HDL-C | −0.0185 | 0.6390 | 0.0106 | 0.7720 | −0.0133 | 0.7378 |
| HDL Size | −0.0175 | 0.6598 | −0.0213 | 0.5661 | −0.0234 | 0.5615 |
Linear regression derived beta (β) values
Triglyceride values were log-transformed for analysis
Table after adjusting for age, race, body-mass index, systolic blood pressure, heart rate, glucose, smoking and alcohol Intake. Cut-off p-value of p<0.01 was used for determining significance.
Unit for VLDL-P and LDL-P was nmol/L, unit for HDL-P was μmol/L
cfPWV = carotid-femoral pulse wave velocity, baPWV = brachial-ankle pulse wave velocity, faPWV = femoral-ankle pulse wave velocity
LDL-P = low-density lipoprotein particle number, HDL-P = high-density lipoprotein particle number, VLDL-P = very low-density lipoprotein particle number
We found no significant interaction between race and VLDL-P, or any of its subclasses in regression models for PWVs. Similarly, interaction terms with race were insignificant for LDL-P, LDL-P subclasses, HDL-P, and HDL-P subclasses except interactions between race and large LDL-P in model for faPWV (p=0.031), between race and HDL-P in model for cfPWV (p=0.049), and between race and HDL-P in the regression model for baPWV (p=0.005). (See Supplementary Table 1)
Discussion
This is the first population-based study to examine the association between NMR lipoprotein subclasses and different measures of arterial stiffness. In the present study, baPWV was significantly associated with lipoprotein subclasses independent of traditional cardiovascular risk factors. Similarly faPWV, a measure of peripheral PWV, was significantly associated with many lipoprotein subclasses after adjusting for traditional risk factors. However, the significant correlation between cfPWV and lipoprotein subclasses became insignificant after adjusting for potential confounders in the regression analysis. The results might suggest that serum lipoproteins play a role in the pathogenesis of arterial stiffness in the peripheral muscular arteries, but not in the aorta among middle-aged men.
In our study, as well as in previous literature,16, 17 cfPWV was not significantly associated with serum lipids, i.e. LDL-C, HDL-C, or triglycerides. Additionally, we did not find significant associations of cfPWV with any of the NMR lipoprotein subclasses. In a systematic-review on the association of cfPWV with CVD risk factors, most studies did not show any significant association between cfPWV and traditional measures of blood lipids i.e. LDL-C, HDL-C, or triglycerides.9 On the other hand, in some previous studies cfPWV is associated with atherosclerotic plaque.18, 19 One plausible explanation may be that calcification of the atherosclerotic plaque increases arterial stiffness. To support this hypothesis, we and others have previously reported significant association between cfPWV and aortic calcification, even as aortic calcification was not associated with traditional serum lipid measures i.e. LDL-C and HDL-C.20–22 More research is needed to understand the pathogenesis of central arterial stiffness, and its association with serum lipids.
We found that baPWV was significantly associated with atherogenic serum lipoproteins, i.e. small LDL-P, independent of other risk factors. Takahashi et al. have also reported a significant association between very small LDL-C and baPWV among men with impaired glucose tolerance, even as lipoprotein subclasses were measured using high performance liquid chromatography in their study, a technique which provides lipoprotein subclass measurement in serum but does not provide lipoprotein particle numbers.23 A possible reason for the association between atherogenic serum lipoproteins and baPWV might be that baPWV, in addition to central arterial stiffness, is also influenced by peripheral arterial atherosclerosis and its related pathophysiological conditions. Consistent with the opposite direction of association for small LDL and large LDL, we found that increasing LDL size was inversely associated with baPWV. Our result is consistent with findings of Paynter et al as they also found lower incidence of hypertension among women with larger LDL size.24 We also found a significant association between small LDL-P and faPWV which further indicates that PWV within a peripheral muscular arterial tree might be a product of arterial stiffness as well as atherosclerosis.
We found that the associations were stronger for lipoprotein subclasses with baPWV than with faPWV, in spite of baPWV encompassing carotid-femoral region where the associations with lipoprotein subclasses are statistically insignificant. One possible reason for this finding might be the inclusion of another peripheral muscular region in the measurement of baPWV i.e. the carotid-brachial segment. We do not have separate heart-brachial PWV data in this study to allow us measure its association with lipoprotein subclasses.
We did not find any significant associations between arterial stiffness and atherogenic serum lipoproteins in the central arterial tree in our middle-aged male population. In contrast, this association was present in the more muscular peripheral arteries. van Popele et al. found a significant association between arterial stiffness and carotid atherosclerosis in an older population.18 In aging arteries, vascular changes occur via an inflammatory milieu leading to increased arterial stiffness.25 It is possible that the arterial stiffening process exposes the arterial wall intima to the deleterious effects of atherogenic LDL particles, leading to stimulation of the localized atherosclerosis, a process that may appear earlier in the peripheral arteries than in central arteries.
We found that total HDL-P as well as small HDL-P was significantly associated with baPWV and with faPWV. In contrast, large HDL-P had an inverse association with baPWV and with faPWV. The reasoning behind our finding of the positive association of total and small HDL-P with PWV is not known, but might be hidden in yet un-deciphered complex pathobiology of HDL synthesis and mode of action. HDL particles are small and disc-shaped to start with, and gain functionality after becoming spherical through a series of poorly understood processes.26 It is possible that large HDL-P represents these functional spherical HDL particles that are likely atheroprotective. Many studies have reported findings similar to ours. From the Women’s Health Study (WHS), Paynter et al have reported independent positive association of total HDL-P (OR = 1.43 (CI: 1.26, 1.61) quintile 5 vs. quintile 1), medium HDL-P (OR = 1.26 (CI: 1.12, 1.41) quintile 5 vs. quintile 1), small HDL-P (OR = 1.23 (CI: 1.30, 1.50) quintile 5 vs. quintile 1) with incident hypertension. In contrast, yet similar to our results, the association between large HDL-P and incident hypertension was protective (OR = 0.81 (CI: 0.72, 0.91) quintile 5 vs. quintile 1).24 Our findings were different from Otvos et al, as they found small HDL-P to be protective of a CHD event in a trial of gemfibrozil among patients with low HDL-C.27 More research is needed to understand the different types of HDL particles and their relative physiological mechanisms.
We found stronger association for baPWV with total HDL-p than any of its individual components i.e. large HDL-p, medium HDL-p, and small HDL-p. This may be due to differences in the relative distribution of HDL-p components among individuals. In our data, there was a significant inverse correlation of large HDL-p with small HDL-p (ρ= −0.22, p<0.0001). As these are associated with baPWV in opposite directions (Table 4), their sum as in total HDL-p has stronger association with baPWV than either of them.
Several recent studies have shown that the inverse association of HDL-C with CHD events is greatly attenuated by adjusting for correlated levels of triglycerides and atherogenic lipoprotein particles measured by apolipoprotein B100 (apoB) or LDL-P,28, 29 but that inverse associations of concentrations of total HDL-P and medium HDL-P with CHD are relatively independent of triglycerides, LDL-P or apoB.28, 30 When we further adjusted for LDL-P in our analysis for association between HDL-P subclasses and PWVs, the results were essentially unchanged thus indicating that the association between HDL-P and its subclasses and PWVs is independent of traditional risk factors and also LDL-P. (See Supplementary Table 2)
We found strong associations for each of large and medium VLDL-p with baPWV as well as faPWV, but not with cfPWV. VLDL particles are secreted by the liver as large, triglyceride-rich particles, deliver fatty acids to peripheral tissues, and become progressively smaller in the circulation as they are depleted of triglycerides. Their direct role in pathogenesis of atherosclerosis or arterial stiffness is not known.31 It is likely that individuals with large number of VLDL-p also have higher LDL-p, thus explaining association between VLDL-p components and, baPWV and faPWV. In our study, the unadjusted Pearson’s correlations of small LDL-p and small HDL-p were significant with large VLDL-p (ρ=0.47 and ρ=0.34 respectively, both p<0.0001), medium VLDL-p (ρ=0.55 and ρ=0.39 respectively, both p<0.0001), and with total VLDL-p (ρ=0.52 and ρ=0.40 respectively, both p<0.0001) (data not shown). Further research is needed to understand the role played by VLDL particles in pathogenesis of arterial stiffness.
We found no significant association between total LDL-P and any of the measures of PWV, independent of traditional risk factors. One likely reason for this unexpected finding might be that the major components of total LDL-P i.e. large LDL-P and small LDL-P are significantly associated with baPWV and faPWV, but in opposite directions. Combining the two lipoprotein subclasses probably obliterated these two associations. In the WHS study, incident hypertension was also inversely associated with large LDL-P and positively associated with small LDL-P, but also with total LDL-P.24 In a supplementary analysis where we additionally adjusted for LDL-C in the regression analysis, we found significant positive association of total LDL-P with baPWV (p=0.004), and with faPWV (p= 0.003), but only weak association with cfPWV (p=0.018) (data not shown). LDL-C is more strongly correlated with large LDL-P (Spearman’s ρ=0.50, p<0.0001) than with small LDL-P (Spearman’s ρ=0.28, p<0.0001) in our study (data not shown) as would be expected as large LDL particles are likely to have larger amount of cholesterol per particle than small LDL particles, and so have stronger correlation with LDL-C. Thus, adjusting for LDL-C possibly adjusted more for the effect of large LDL-P than effect of small LDL-P, resulting in a significant positive association of total LDL-P with each of baPWV and faPWV.
We found lower cfPWV and baPWV among Koreans than any other racial group. Also, Japanese Americans had higher baPWV and faPWV than the other two groups. However, we combined the population groups for the analysis as we did not find significant interaction by race for most of the lipoprotein subclasses (Supplementary Table 1).
Our study had several limitations. As this is a cross-sectional study, we cannot establish temporality of the association between lipoprotein particles and PWV. We had only cross-sectional data for analysis, thus we can’t comment on causality of increased PWV. Our study population comprised of men within a relatively narrow age range, so our results cannot be extrapolated to women or other age groups. However, the narrow age range likely provided us with more precision in detecting the association between the two sets of CVD markers. We excluded individuals with diabetes from this analysis. The results, however, remained unchanged after including them in a supplementary analysis (data not shown). By keeping the cut-off p-value at 0.01, we might have missed potentially important associations between the two sets of biomarkers. However, keeping α at 0.01 likely prevented us from detecting false associations that occur by chance.
Conclusion
In summary, our results demonstrate that among apparently healthy middle-aged men, NMR spectroscopy-derived lipoprotein subclasses are associated with baPWV and faPWV, but not with cfPWV. While small LDL-P and small HDL-P have significant positive associations with baPWV and faPWV, the association of each of large LDL-P and large HDL-P with each of baPWV and faPWV are in inverse direction. In contrast, traditionally measured VLDL-C, LDL-C and HDL-C do not have significant associations with any measure of arterial stiffness i.e. cfPWV, baPWV or faPWV. More research is needed to understand the differences in the pathogenesis of central vs. peripheral arterial stiffness.
Supplementary Material
Summary Table.
What is known about this topic
|
What this study adds
|
Acknowledgments
This research was funded by the Korea Centers for Disease Control and Prevention, Ministry for Health and Welfare 2004-E71001–00, 2005-E71001-00, Korea, R01HL68200, R01HL071561, and N01AG12100, from the National Institutes of Health, the USA, B 16790335 and A 13307016 from the Japanese Ministry of Education, Culture, Sports, Science and Technology, Japan. We thank Sara Strollo, MPH for her editorial assistance during preparation of this manuscript.
Footnotes
Conflict of Interest:
Rachel Mackey has been PI of an unrestricted research grant from LipoScience, Inc. to the University of Pittsburgh. LipoScience had no involvement in the design/analysis/conduct of that study or this study and manuscript other than performing the blinded laboratory measurements. Other authors declare no potential conflict of interest.
Supplementary information is available at JHH’s website
Disclosure:
This research was funded by the Korea Centers for Disease Control and Prevention, Ministry for Health and Welfare 2004-E71001–00, 2005-E71001-00, Korea, R01HL68200, R01HL071561, and N01AG12100, from the National Institutes of Health, the USA, B 16790335 and A 13307016 from the Japanese Ministry of Education, Culture, Sports, Science and Technology, Japan.
References
- 1.Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C, Hayoz D, et al. Expert consensus document on arterial stiffness: Methodological issues and clinical applications. Eur Heart J. 2006;27:2588–605. doi: 10.1093/eurheartj/ehl254. [DOI] [PubMed] [Google Scholar]
- 2.Vlachopoulos C, Aznaouridis K, Stefanadis C. Prediction of cardiovascular events and all-cause mortality with arterial stiffness: A systematic review and meta-analysis. J Am Coll Cardiol. 2010;55:1318–27. doi: 10.1016/j.jacc.2009.10.061. [DOI] [PubMed] [Google Scholar]
- 3.Mancia G, De Backer G, Dominiczak A, Cifkova R, Fagard R, Germano G, et al. 2007 guidelines for the management of arterial hypertension: The task force for the management of arterial hypertension of the european society of hypertension (esh) and of the european society of cardiology (esc) Eur Heart J. 2007;28:1462–536. doi: 10.1093/eurheartj/ehm236. [DOI] [PubMed] [Google Scholar]
- 4.Tanaka H, Munakata M, Kawano Y, Ohishi M, Shoji T, Sugawara J, et al. Comparison between carotid-femoral and brachial-ankle pulse wave velocity as measures of arterial stiffness. J Hypertens. 2009;27:2022–7. doi: 10.1097/HJH.0b013e32832e94e7. [DOI] [PubMed] [Google Scholar]
- 5.Vlachopoulos C, Aznaouridis K, Terentes-Printzios D, Ioakeimidis N, Stefanadis C. Prediction of cardiovascular events and all-cause mortality with brachial-ankle elasticity index: A systematic review and meta-analysis. Hypertension. 2012;60:556–62. doi: 10.1161/HYPERTENSIONAHA.112.194779. [DOI] [PubMed] [Google Scholar]
- 6.Sugawara J, Hayashi K, Yokoi T, Cortez-Cooper MY, DeVan AE, Anton MA, Tanaka H. Brachial-ankle pulse wave velocity: An index of central arterial stiffness? J Hum Hypertens. 2005;19:401–6. doi: 10.1038/sj.jhh.1001838. [DOI] [PubMed] [Google Scholar]
- 7.Arsenault BJ, Boekholdt SM, Kastelein JJ. Lipid parameters for measuring risk of cardiovascular disease. Nat Rev Cardiol. 2011;8:197–206. doi: 10.1038/nrcardio.2010.223. [DOI] [PubMed] [Google Scholar]
- 8.Otvos JD, Mora S, Shalaurova I, Greenland P, Mackey RH, Goff DC., Jr Clinical implications of discordance between low-density lipoprotein cholesterol and particle number. J Clin Lipidol. 2011;5:105–13. doi: 10.1016/j.jacl.2011.02.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cecelja M, Chowienczyk P. Dissociation of aortic pulse wave velocity with risk factors for cardiovascular disease other than hypertension: A systematic review. Hypertension. 2009;54:1328–36. doi: 10.1161/HYPERTENSIONAHA.109.137653. [DOI] [PubMed] [Google Scholar]
- 10.Sekikawa A, Curb JD, Ueshima H, El-Saed A, Kadowaki T, Abbott RD, et al. Marine-derived n-3 fatty acids and atherosclerosis in japanese, japanese-american, and white men: A cross-sectional study. J Am Coll Cardiol. 2008;52:417–24. doi: 10.1016/j.jacc.2008.03.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Abbott RD, Ueshima H, Rodriguez BL, Kadowaki T, Masaki KH, Willcox BJ, et al. Coronary artery calcification in japanese men in japan and hawaii. Am J Epidemiol. 2007;166:1280–7. doi: 10.1093/aje/kwm201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sekikawa A, Ueshima H, Kadowaki T, El-Saed A, Okamura T, Takamiya T, et al. Less subclinical atherosclerosis in japanese men in japan than in white men in the united states in the post-world war ii birth cohort. Am J Epidemiol. 2007;165:617–24. doi: 10.1093/aje/kwk053. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sekikawa A, Ueshima H, Sutton-Tyrrell K, Kadowaki T, El-Saed A, Okamura T, et al. Intima-media thickness of the carotid artery and the distribution of lipoprotein subclasses in men aged 40 to 49 years between whites in the united states and the japanese in japan for the era jump study. Metabolism. 2008;57:177–82. doi: 10.1016/j.metabol.2007.08.022. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Otvos JD. Measurement of lipoprotein subclass profiles by nuclear magnetic resonance spectroscopy. Clin Lab. 2002;48:171–80. [PubMed] [Google Scholar]
- 15.Kimoto E, Shoji T, Shinohara K, Hatsuda S, Mori K, Fukumoto S, et al. Regional arterial stiffness in patients with type 2 diabetes and chronic kidney disease. J Am Soc Nephrol. 2006;17:2245–52. doi: 10.1681/ASN.2005101038. [DOI] [PubMed] [Google Scholar]
- 16.Hansen TW, Jeppesen J, Rasmussen S, Ibsen H, Torp-Pedersen C. Relation between insulin and aortic stiffness: A population-based study. J Hum Hypertens. 2004;18:1–7. doi: 10.1038/sj.jhh.1001620. [DOI] [PubMed] [Google Scholar]
- 17.Nguyen QM, Srinivasan SR, Xu JH, Chen W, Berenson GS. Racial (black-white) divergence in the association between adiponectin and arterial stiffness in asymptomatic young adults: The bogalusa heart study. Am J Hypertens. 2008;21:553–7. doi: 10.1038/ajh.2008.14. [DOI] [PubMed] [Google Scholar]
- 18.van Popele NM, Grobbee DE, Bots ML, Asmar R, Topouchian J, Reneman RS, et al. Association between arterial stiffness and atherosclerosis: The rotterdam study. Stroke. 2001;32:454–60. doi: 10.1161/01.str.32.2.454. [DOI] [PubMed] [Google Scholar]
- 19.Zureik M, Temmar M, Adamopoulos C, Bureau JM, Courbon D, Thomas F, et al. Carotid plaques, but not common carotid intima-media thickness, are independently associated with aortic stiffness. J Hypertens. 2002;20:85–93. doi: 10.1097/00004872-200201000-00013. [DOI] [PubMed] [Google Scholar]
- 20.Sekikawa A, Shin C, Curb JD, Barinas-Mitchell E, Masaki K, El-Saed A, et al. Aortic stiffness and calcification in men in a population-based international study. Atherosclerosis. 2012;222:473–7. doi: 10.1016/j.atherosclerosis.2012.03.027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Cecelja M, Jiang B, Bevan L, Frost ML, Spector TD, Chowienczyk PJ. Arterial stiffening relates to arterial calcification but not to noncalcified atheroma in women. A twin study. J Am Coll Cardiol. 2011;57:1480–6. doi: 10.1016/j.jacc.2010.09.079. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.McEniery CM, McDonnell BJ, So A, Aitken S, Bolton CE, Munnery M, et al. Aortic calcification is associated with aortic stiffness and isolated systolic hypertension in healthy individuals. Hypertension. 2009;53:524–31. doi: 10.1161/HYPERTENSIONAHA.108.126615. [DOI] [PubMed] [Google Scholar]
- 23.Takahashi R, Imamura A, Yoshikane M, Suzuki M, Murakami R, Cheng XW, et al. Very small low-density lipoprotein cholesterol level is a determinant of arterial stiffness in men with impaired glucose metabolism. J Atheroscler Thromb. 2010;17:1282–9. doi: 10.5551/jat.5272. [DOI] [PubMed] [Google Scholar]
- 24.Paynter NP, Sesso HD, Conen D, Otvos JD, Mora S. Lipoprotein subclass abnormalities and incident hypertension in initially healthy women. Clin Chem. 2011;57:1178–87. doi: 10.1373/clinchem.2011.167544. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Johnson CP, Baugh R, Wilson CA, Burns J. Age related changes in the tunica media of the vertebral artery: Implications for the assessment of vessels injured by trauma. J Clin Pathol. 2001;54:139–45. doi: 10.1136/jcp.54.2.139. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Tsompanidi EM, Brinkmeier MS, Fotiadou EH, Giakoumi SM, Kypreos KE. Hdl biogenesis and functions: Role of hdl quality and quantity in atherosclerosis. Atherosclerosis. 2010;208:3–9. doi: 10.1016/j.atherosclerosis.2009.05.034. [DOI] [PubMed] [Google Scholar]
- 27.Otvos JD, Collins D, Freedman DS, Shalaurova I, Schaefer EJ, McNamara JR, et al. Low-density lipoprotein and high-density lipoprotein particle subclasses predict coronary events and are favorably changed by gemfibrozil therapy in the veterans affairs high-density lipoprotein intervention trial. Circulation. 2006;113:1556–63. doi: 10.1161/CIRCULATIONAHA.105.565135. [DOI] [PubMed] [Google Scholar]
- 28.Mackey RH, Greenland P, Goff DC, Jr, Lloyd-Jones D, Sibley CT, Mora S. High-density lipoprotein cholesterol and particle concentrations, carotid atherosclerosis, and coronary events: Mesa (multi-ethnic study of atherosclerosis) J Am Coll Cardiol. 2012;60:508–16. doi: 10.1016/j.jacc.2012.03.060. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.El Harchaoui K, Arsenault BJ, Franssen R, Despres JP, Hovingh GK, Stroes ES, et al. High-density lipoprotein particle size and concentration and coronary risk. Ann Intern Med. 2009;150:84–93. doi: 10.7326/0003-4819-150-2-200901200-00006. [DOI] [PubMed] [Google Scholar]
- 30.van der Steeg WA, Holme I, Boekholdt SM, Larsen ML, Lindahl C, Stroes ES, et al. High-density lipoprotein cholesterol, high-density lipoprotein particle size, and apolipoprotein a-i: Significance for cardiovascular risk: The ideal and epic-norfolk studies. J Am Coll Cardiol. 2008;51:634–42. doi: 10.1016/j.jacc.2007.09.060. [DOI] [PubMed] [Google Scholar]
- 31.Taskinen MR. Diabetic dyslipidaemia: From basic research to clinical practice. Diabetologia. 2003;46:733–49. doi: 10.1007/s00125-003-1111-y. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.