Lipoprotein Subclasses in the Monitored Atherosclerosis Regression Study (MARS)
Treatment Effects and Relation to Coronary Angiographic Progression
Abstract Accumulating evidence suggests that triglyceride-rich lipoproteins contribute to coronary artery disease. Using data from the Monitored Atherosclerosis Regression Study, an angiographic trial of middle-aged men and women randomized to lovastatin or placebo, we investigated relationships between lipoprotein subclasses and progression of coronary artery atherosclerosis. Coronary artery lesion progression was determined by quantitative coronary angiography in low-grade (<50% diameter stenosis), high-grade (≥50% diameter stenosis), and all coronary artery lesions in 220 baseline/2-year angiogram pairs. Analytical ultracentrifugation was used to measure lipoprotein masses that were statistically evaluated for treatment group differences and relationships to progression of coronary artery atherosclerosis. All low density lipoprotein (LDL), intermediate density lipoprotein (IDL), and very low density lipoprotein (VLDL) masses were significantly lowered and all high density lipoprotein (HDL) masses were significantly raised with lovastatin therapy. The mass of smallest LDL (Svedberg flotation rate [Sf] 0 to 3), IDL (Sf 12 to 20), all VLDL subclasses (Sf 20 to 60, Sf 60 to 100, and Sf 100 to 400), and peak LDL flotation rate were significantly related to the progression of coronary artery lesions, specifically low-grade lesions. Greater baseline levels of HDL3 were related to a lower likelihood of coronary artery lesion progression. In multivariate analyses, small VLDL (Sf 20 to 60) and HDL3 mass were the most important correlates of coronary artery lesion progression. These results provide further evidence for the importance of triglyceride-rich lipoproteins in the progression of coronary artery disease. In addition, these results present new evidence for the possible protective role of HDL3 in the progression of coronary artery lesions. More specific information on coronary artery lesion progression may be obtained through the study of specific apolipoprotein B–containing lipoproteins.
- triglyceride-rich lipoproteins
- coronary angiography
- angiographic trials
- lipoprotein subclasses
- coronary artery disease
Reprint requests to Howard N. Hodis, MD, Atherosclerosis Research Unit, Division of Cardiology, University of Southern California School of Medicine, 2250 Alcazar St, CSC 132, Los Angeles, CA 90033.
- Received August 17, 1995.
- Accepted January 3, 1996.
The Monitored Atherosclerosis Regression Study was a double-blind, placebo-controlled, randomized serial angiographic trial testing whether reduction of LDL cholesterol by lovastatin would favorably alter the progression of coronary artery atherosclerosis.1 The primary study end point was average change in percent diameter stenosis of coronary artery lesions measured by QCA. Study results indicated no significant treatment effect for all lesions (average Δ%S=+2.2 in placebo, +1.6 in lovastatin group; NS). A significant treatment effect was seen for high-grade (≥50%S) baseline lesions (average Δ%S=+0.9 in placebo, −4.1 in lovastatin groups; P<.01) but not for low-grade (<50%S) baseline lesions (average Δ%S=+3.0 in placebo, +2.6 in lovastatin group; NS).2
We have previously reported a within-treatment group analysis of clinical and lipid risk factors for angiographic progression of CAD in MARS.3 In the placebo group, risk factors for the progression of low-grade (<50%S) lesions were total triglycerides and the total cholesterol/HDL cholesterol ratio. Risk factors for progression of high-grade (≥50%S) lesions were HDL cholesterol, and the LDL cholesterol/HDL cholesterol and total cholesterol/HDL cholesterol ratios. In the lovastatin group, risk factors for progression of low-grade lesions were total triglycerides and the VLDL-LDL–associated apo C-III–heparin precipitate, a marker of triglyceride-rich lipoprotein metabolism.4 Risk factors for progression of high-grade lesions were LDL cholesterol, apolipoprotein B, and the LDL cholesterol/HDL cholesterol and total cholesterol/HDL cholesterol ratios. These results indicate that triglyceride-rich lipoproteins (and their markers) and cholesterol ester–rich lipoproteins (and their markers) have differential effects on coronary artery lesion progression according to lesion severity.
Reports from other serial coronary angiographic studies have provided additional evidence that triglyceride-rich lipoproteins play an important role in the progression of CAD.5 To further investigate the role of triglyceride-rich lipoproteins in CAD, we now extend these analyses to investigate the relationship between coronary artery lesion progression and mass of lipoprotein subclasses measured by analytical ultracentrifugation. We also report the effects of lovastatin on these lipoprotein subclasses. We hypothesized that triglyceride-rich lipoprotein subclasses would correlate with coronary artery lesion progression. We further hypothesized that consistent with our earlier report,3 triglyceride-rich lipoprotein subclasses would show a greater correlation with progression of low-grade versus high-grade coronary artery lesions.
MARS Study Design
Subjects were 270 men and women, 37 to 67 years of age, with total cholesterol levels ranging from 4.92 to 7.64 mmol/L (190 to 295 mg/dL). All subjects had CAD documented by coronary angiography, meeting the following criteria: (1) at least two coronary segments exhibiting CAD and (2) at least one unoccluded segment showing a minimum 50%S unaltered by percutaneous transluminal coronary angioplasty. Subjects were randomized to lovastatin 80 mg daily or lovastatin placebo. All subjects had dietary counseling with dietary cholesterol goals of 250 mg or less per day, total dietary fat 27% of calories or less, saturated fat 7% of calories or less, and monounsaturated and polyunsaturated dietary fats each with 10% of calories or less.1
Coronary Angiography and Study End Points
Baseline/2-year coronary angiographic film pairs were obtained on 247 study subjects by use of the percutaneous femoral technique with sufficient right and left anterior oblique views to clearly visualize all coronary artery lesions. After completion of the 2-year angiogram, QCA was performed blinded to treatment group assignment. Baseline/2-year film pairs were processed in tandem with dual projectors to match frames for orientation and degree of contrast filling. Percent diameter stenosis of all coronary artery lesions identified by a panel of expert angiographers was measured.1 Three sequential frames exposed during end diastole were digitized when possible; otherwise three sequential frames from other phases of the cardiac cycle were used. For each lesion evaluable by QCA, %S (at baseline and follow-up) and Δ%S (equal to follow-up minus baseline %S) were measured and averaged over the three frames.6 The primary study end point was lesion Δ%S, averaged over all QCA-evaluable lesions. The Δ%S was also separately averaged over low-grade (<50%S) and high-grade (≥50%S) baseline lesions. Twenty-three film pairs with nitroglycerin administered to the subject during one of the angiograms only and four film pairs not evaluable by QCA were excluded from the analysis, leaving 220 film pairs with QCA data.
Plasma samples for lipoprotein subclasses were obtained from study subjects after an 8-hour fast. Samples were stored at −4°C and transported on wet ice to the Donner Laboratory, Lawrence Berkeley National Laboratory, University of California, Berkeley. Total mass of lipoprotein subclasses was measured by analytical ultracentrifugation, which generates a Schlieren curve representing the distribution of lipoprotein particle mass as a function of particle flotation rate (Sf). Lipoproteins of flotation rate Sf 0 to 400 were measured in the d<1.063 g/mL fraction of plasma separated by preparative ultracentrifugation.7 Computer-based analysis of the resulting Schlieren curves then was used to measure the total mass of VLDL in 14 intervals between Sf 20 to 400, IDL in four intervals between Sf 12 to 20, and LDL in 11 intervals between Sf 0 to 12. A measure of the peak LDL flotation rate was also obtained as the flotation rate at which LDL mass concentration was greatest. HDLs of flotation rate F1.20 0 to 9 were measured in the d<1.21 g/mL fraction of plasma.7 Masses in specific intervals were summed to obtain measures of total lipoprotein mass of three HDL subclasses (HDL3 [F1.20 0 to 3.5], HDL2 [F1.20 3.5 to 9], and total HDL [F1.20 0 to 9]), five LDL subclasses (LDL-IV [Sf 0 to 3], LDL-III [Sf 3 to 5], LDL-II [Sf 5 to 7], LDL-I [Sf 7 to 12]), and total LDL [Sf 0 to 12]), one IDL subclass (Sf 12 to 20), and four VLDL subclasses (small VLDL [Sf 20 to 60], intermediate VLDL [Sf 60 to 100], large VLDL [Sf 100 to 400], and total VLDL [Sf 20 to 400]). Samples for measurements of lipoprotein subclasses were obtained at one baseline visit and four on-trial visits (every 6 months on-trial). Of the 270 randomized subjects, 263 had baseline, 264 had on-trial, and 257 had both baseline and on-trial lipoprotein subclass measurements.
Treatment group comparisons of baseline levels and on-trial changes in lipoprotein subclasses were performed with the nonparametric Wilcoxon rank sum test because of the nonnormality of many of these measures. For the same reason, within-group changes from baseline in lipoprotein subclasses were analyzed with the Wilcoxon signed rank test.
Logistic regression procedures were used to determine whether lipoprotein subclasses were correlated with angiographic progression of coronary artery lesions. The following binary indicators of progression were used: (1) progression/no progression (Δ%S>0) based on average Δ%S in low-grade (<50%S) lesions and (2) progression/no progression (Δ%S>0) based on average change in high-grade (≥50%S) lesions. Since other studies examining angiographic relationships to these types of data have evaluated progression over all coronary artery lesions, we also examined associations with a binary progression variable of progression/no progression (Δ%S>0) based on average change in all evaluable coronary artery lesions. The cutpoint at 0 was chosen to represent an overall average progression (average change >0) or nonprogression (average change ≤0) and was the only cutpoint attempted. This cutpoint is significantly associated with an assessment of overall coronary angiographic change derived from consensus panels of expert angiographers (the global coronary change score ) when it is similarly dichotomized to represent overall coronary progression versus nonprogression (P<.0001 for the QCA cutpoint evaluated over all, low-grade, and high-grade lesions).
Logistic regression analyses were conducted within each treatment group and in the combined sample with a covariate included to control for treatment group. Baseline and on-trial levels of the lipoprotein subclasses were tested for their association with these binary progression variables. Tests of association of on-trial levels of lipoprotein subclasses with coronary artery lesion progression were adjusted for baseline levels of the lipoprotein subclass. ORs and 95% CIs were computed and expressed per approximate standard deviation (computed in the total sample) of each lipoprotein subclass. Likelihood ratio χ2 tests were used to compute significance levels on the ORs, and a two-sided α of 0.05 was used. As a corroborative analysis, multiple linear regression methods were also used, using Δ%S as a continuous dependent variable. Because results were essentially the same as provided by the logistic regression analyses, and OR estimates provided a more easily interpretable measure of association, the results of the logistic regression analyses are presented as the primary analysis.
Multivariate logistic regression models were developed with the use of a stepwise forward selection process. At the first step, the on-trial lipoprotein subclass with the largest likelihood ratio χ2 and its corresponding baseline level were included in the model. Subsequent steps tested the statistical significance of the additional contribution of other on-trial lipoprotein subclasses (with their baseline level forced into the model). This process was repeated until no variables added significantly to the multivariate model.
Clinical features as well as average lipid and apolipoprotein levels in the 220 subjects with both QCA end point and lipoprotein subclass data are provided in Table 1⇓. The majority of subjects were men, with an average age of 58 years. Although the cohort on average was normotensive at randomization, 45% (98 subjects) had a history of hypertension. The large majority of subjects (82%) were nonsmokers at baseline; however, many of these subjects (61% of the sample) were former smokers. More than 50% of the subjects had had a myocardial infarction and approximately 15% of subjects had undergone a coronary revascularization procedure (angioplasty or bypass surgery) at some time before the study.
Thirty-four percent (n=74) of the subjects had total cholesterol levels of 6.20 mmol/L (240 mg/dL) or greater, and 44% (n=96) had LDL cholesterol levels of 4.1 mmol/L (160 mg/dL) or greater at baseline. Treatment group comparisons on on-trial changes in lipids and apolipoproteins have been described in detail elsewhere.2 3 In brief, the lovastatin group exhibited significant decreases in total cholesterol, LDL cholesterol, total triglyceride, and apolipoprotein B levels and a significant increase in HDL cholesterol.
Baseline/2-year angiographic film pairs were evaluated by QCA for 1889 total lesions in the 220 subjects (an average of 8.6 lesions per subject). The majority of lesions (83%, 1569 lesions) were <50%S at baseline. On average, lesions were moderately stenotic at baseline (Table 1⇑). Angiographic progression of lesions <50%S (average Δ%S>0) was evident in 132 of 217 subjects (61%), while angiographic progression of lesions ≥50%S was evident in 57 of 156 subjects (37%).
Lipoprotein Subclasses: Treatment Effects
Consistent with lipid and apolipoprotein changes previously reported in the MARS study,2 3 the lovastatin-treated group showed statistically significant on-trial decreases in all LDL, IDL, and VLDL subclasses relative to the placebo group (Table 2⇓). Lipoprotein mass in HDL2, and to a lesser extent HDL3, showed significant on-trial increases in the lovastatin group. The peak LDL flotation rate also showed a significant increase in the lovastatin group relative to the placebo group, indicating a shift to larger, more buoyant LDL particles.
Lipoprotein Subclasses: Relationship to Coronary Artery Lesion Progression
Table 3⇓ displays the significant correlates of angiographic progression (average Δ%S>0) analyzed within each treatment group and for the combined sample. Analyses of all on-trial levels of lipoprotein subclasses are adjusted for their baseline levels. OR estimates are expressed per approximate standard deviation (computed in the total sample) of each lipoprotein subclass. OR estimates >1 indicate that higher levels of the variable are associated with angiographic progression, and OR estimates <1 indicate that higher levels of the variable are associated with nonprogression. OR estimates significantly different from 1 are highlighted in Table 3⇓; for comparison, other analyses for the significant lipoprotein subclasses are also presented.
Progression of High-Grade (≥50%S) Lesions
No lipoprotein subclass variables were significantly related to the progression of high-grade (≥50%S) lesions in either the placebo, lovastatin, or combined treatment groups. For comparison with OR estimates associated with progression of low-grade and all coronary artery lesions, the results for these high-grade coronary artery lesions are provided for the combined treatment sample in Table 3⇑. All OR estimates were close to the null value of 1.0, ranging from 0.8 (baseline HDL3) to 1.3 (on-trial peak LDL flotation rate), in contrast to statistically significantly elevated OR estimates for other lesion classes, ranging from 1.7 to 2.9. Similar results were obtained for analyses conducted by treatment group.
Correlates of Coronary Artery Lesion Progression in the Placebo Group
In the placebo group, higher on-trial levels of LDL-IV (Sf 0 to 3), IDL, small VLDL, intermediate VLDL, large VLDL, and total VLDL masses were correlated with progression of low-grade (<50%S) baseline lesions. These results were essentially equivalent when progression was evaluated over all coronary artery lesions in the placebo group; however, LDL-IV and large VLDL were not significantly associated to progression over all coronary artery lesions. Similar to LDL-IV, the OR estimates for LDL-III (Sf 3 to 5) were also elevated (in the placebo group, OR=1.5 per 30 mg/dL, 95% CI=0.9 to 2.5 for progression of low-grade lesions and OR=1.4 per 30 mg/dL, 95% CI=0.8 to 2.2 for progression evaluated over all coronary artery lesions); however, these ORs were not statistically significantly elevated. ORs for LDL-II (Sf 5 to 7) and LDL-I (Sf 7 to 12) were not elevated (data not shown).
Correlates of Coronary Artery Lesion Progression in the Lovastatin Group
In the lovastatin group, the single statistically significant correlate of coronary angiographic progression of low-grade lesions was a lower on-trial peak LDL flotation rate. Although not statistically significant, OR estimates for all VLDL subclasses were elevated and roughly equivalent to corresponding ORs in the placebo group. When angiographic progression was evaluated over all coronary lesions, the single significant correlate of progression was a lower level of HDL3 mass at baseline.
Correlates of Coronary Artery Lesion Progression in the Combined Treatment Groups
In the combined sample, analyses were conducted controlling for treatment group as well as baseline levels of lipoprotein subclasses (Table 3⇑). Significant correlates of progression of low-grade lesions were higher on-trial levels of IDL, small and intermediate VLDL, and total VLDL masses and a lower on-trial peak LDL flotation rate. When evaluated over all coronary artery lesions, lower levels of baseline HDL3 and higher on-trial levels of LDL-IV and small, intermediate, and total VLDL masses were significant correlates of coronary artery lesion progression.
Stepwise Multivariate Models
Stepwise logistic regression procedures were used to identify statistically independent correlates of coronary artery lesion progression. For each model, independent variables selected for possible entry into the stepwise model were those lipoprotein subclasses identified as significant correlates in Table 3⇑. Stepwise procedures were conducted for the placebo and combined treatment groups. Results of these multivariate models are shown in Table 4⇓.
In the placebo group, the on-trial level of small VLDL mass was the single independent correlate of progression of low-grade lesions. The on-trial levels of IDL and intermediate VLDL masses were statistically independent correlates of angiographic progression evaluated over all coronary artery lesions.
Because only single lipoprotein subclasses were significantly related to progression of coronary disease in the lovastatin group, the multivariate results are equivalent to Table 3⇑ results. Specifically, the on-trial peak LDL flotation rate was inversely related to progression of low-grade coronary artery lesions, and the HDL3 mass at baseline was inversely related to progression evaluated over all coronary artery lesions.
In the combined treatment groups, the stepwise model forced in treatment group as well as baseline levels of significant lipoprotein subclasses. The on-trial small VLDL mass was the single independent correlate of coronary artery lesion progression of low-grade lesions. Statistically independent correlates of progression evaluated over all coronary artery lesions in the stepwise model were baseline HDL3 and the on-trial small VLDL mass.
Multiple Linear Regression Models
Coronary angiographic progression was also modeled as a continuous outcome variable. Because results in general were markedly similar to those obtained using logistic regression procedures in Table 3⇑, only results for the combined sample, controlling for baseline lipoprotein subclasses and treatment group, are provided. Lipoprotein subclasses significantly positively correlated to angiographic progression evaluated over all coronary artery lesions (Δ%S) were on-trial levels of LDL-IV (partial r=.17, P<.02), IDL (partial r=.15, P<.03), and small VLDL (partial r=.28), intermediate VLDL (partial r=.42), large VLDL (partial r=.41), and total VLDL (partial r=.41, all P<.0001). Negative correlations with baseline HDL3 (partial r=−.13, P=.053) and on-trial peak LDL flotation rate (partial r=−.13, P=.052) were of marginal significance. When progression of low-grade coronary artery lesions was modeled as a continuous variable, lipoprotein subclasses significantly correlated to progression were on-trial levels of IDL (partial r=.17, P<.02) and small VLDL (partial r=.39), intermediate VLDL (partial r=.41), large VLDL (partial r=.36), and total VLDL (partial r=.41, all P<.0001). On-trial levels of LDL-II (partial r=−.16, P<.02) were negatively correlated to progression of low-grade lesions, and the negative correlation with peak LDL flotation rate was of marginal significance (partial r=−.13, P=.07). After adjustment for baseline subclass level and treatment group, no subclasses were significantly related to the progression of high-grade coronary artery lesions.
It is well established that elevated LDL cholesterol levels are associated with the risk for development of CAD. Coronary angiographic trials have demonstrated that lowering LDL cholesterol levels reduces progression of coronary artery lesions, which in turn is related to a reduction in clinical coronary events.8 9 10 HDL cholesterol levels are inversely related to coronary heart disease risk. However, the relation of HDL cholesterol levels and CHD risk is complicated by the inverse association between triglyceride-rich lipoproteins and HDL cholesterol metabolism. Less well understood is the independent relationship of triglyceride-rich lipoproteins and progression of coronary artery lesions. Nonetheless, there is a growing body of evidence that indicates that triglyceride-rich lipoproteins are atherogenic.5 By examination of specific lipoprotein subclasses, results from this study add compelling new evidence for the independent roles of triglyceride-rich lipoproteins in promoting the progression of CAD and of HDL3 in reducing the progression of CAD.
Lipoprotein Subclass Relationships by Coronary Artery Lesion Severity
Results from this study confirm our previous findings that lipoprotein risk factors have a differential effect on lesion progression according to baseline lesion severity. When examined in lesions <50%S at baseline, coronary angiographic progression was associated primarily with lipoprotein levels in the intermediate (IDL) and very low density (VLDL) range. These results were statistically significant in the placebo and combined samples. Although not statistically significant, ORs corresponding to these lipoprotein subclasses were equivalently elevated in the drug group. Interaction tests indicated that ORs for the VLDL subclasses did not vary by treatment group, suggesting that these relationships to progression of low-grade lesions were apparent throughout a range of lipoprotein levels. The inverse association in the lovastatin-treated group between the LDL peak flotation rate and progression of lesions <50%S is consistent with the association between VLDL mass and progression of these lesions, since LDL size and density are inversely related to triglyceride-rich lipoprotein levels.11
In this study, no lipoprotein subclasses were related to the progression of high-grade baseline lesions (≥50%S). Because measurement of higher-grade coronary artery lesions by QCA entails more measurement error than does measurement of less severe lesions,12 it is possible that measures of association were biased toward the null value of 1, with an associated loss of statistical power. However, given the consistency of these null findings across all lipoprotein subclasses, it is more likely that these results reflect a true finding. Additionally, since cellular fibrous tissue is often found at earlier stages of plaque development (<50%S) and dense fibrous tissue at later stages of plaque development (≥50%S), there may be a pathological explanation for the differential relationship of lipoproteins and lesion progression according to baseline lesion severity.13 14
The large majority of studies that have used coronary angiography to assess the relationship between the extent of existing CAD or coronary artery lesion progression with levels of various risk factors have assessed progression over all coronary artery lesions. For comparison with these reports, we also analyzed the relationship of lipoprotein subclasses to angiographic change in all evaluable coronary lesions. In general, these analyses produced results that were remarkably similar to those for low-grade baseline lesions alone (Table 3⇑). In addition, HDL3 mass (F1.20 0 to 3.5) measured at baseline was found to be significantly inversely associated with coronary artery lesion progression.
Lipoprotein Subclasses and Progression of Coronary Artery Disease
Studies that have cross-sectionally examined the relationship between lipoprotein subclasses and the extent or severity of CAD assessed angiographically have implicated triglyceride-rich lipoproteins (including IDL and VLDL) as major contributors to coronary atherosclerosis.15 16 17 In particular, lipoproteins in the Sf 12 to 60 range (IDL and small VLDL levels) in these studies emerged as independent correlates of the extent of CAD.
Other than this report, there is little information available concerning the relationship between lipoprotein subclasses and progression of coronary artery lesions.5 Krauss et al18 measured lipoprotein mass concentrations by analytical ultracentrifugation at baseline and after 2 years in a subset of 57 men from the NHLBI Type II Coronary Intervention Study by the same methods for measurement of lipoproteins as those in the present study. Lesion progression over 5 years was associated with change in IDL mass (Sf 12 to 20) over 2 years but not with change in LDL mass (Sf 0 to 12), VLDL mass (Sf 20 to 400), HDL2 mass (F1.20 3.5 to 9), or HDL3 mass (F1.20 0 to 3.5). The association of triglyceride-rich lipoproteins with the progression of CAD in the NHLBI Type II Study is consistent with the results of the present study, particularly for lesions <50%S in the placebo group. Since the placebo group in our study was an unintervened group, these results have important implications for the independent role of triglyceride-rich lipoproteins in the natural progression of CAD.
In a more recent report, on-trial concentrations of IDL (δ=1.006 to 1.019 kg/L), LDL2 (δ=1.019 to 1.040 kg/L), LDL3 (δ=1.040 to 1.063 kg/L), and HDL3 (δ=1.125 to 1.210 kg/L) subfractions were correlated with change in arterial luminal dimensions.19 On-trial LDL3 concentration was the most strongly associated lipoprotein with luminal dimensional changes. However, because of limited baseline lipoprotein measures, these on-trial associations with progression of CAD were not adjusted for the possible confounding bias of baseline lipoprotein subclass levels. In another recent study, progression of CAD was found to be related to the sum of a calculated IDL fraction and estimated level of remnant VLDL cholesterol, which was also a statistically significant prognostic factor for the development of clinical coronary events.20
In the present study, when LDL subfraction masses were examined in the intervals Sf 0 to 3, Sf 3 to 5, Sf 5 to 7, and Sf 7 to 12, LDL mass Sf 0 to 3 was significantly correlated with progression of all coronary artery lesions in both placebo and lovastatin groups combined and of low-grade lesions in the placebo group. This suggests that the smallest, most dense LDL particles are associated with progression of coronary artery lesions, particularly low-grade lesions. However, this small LDL-IV mass (Sf 0 to 3) was the LDL subclass least affected by lovastatin therapy (Table 2⇑). The LDL-IV mass was also highly correlated with all VLDL subclasses (r=.49 to .67 in the placebo group, all P<.0001) and specifically, small VLDL (r=.67 in the placebo group, P<.0001), whereas other LDL subclasses were less strongly or inversely correlated with the VLDL subclasses. Small, dense LDL particles are the major LDL species characteristic of the atherogenic pattern B LDL phenotype. This phenotype has been further characterized by increased levels of triglycerides, IDL, and VLDL masses.21 Elevated levels of this small LDL mass thus may be a marker for elevated levels of triglyceride-rich lipoproteins, particularly VLDLs.
In the same cohort of MARS subjects reported here, we previously demonstrated that apo C-III was a predominant risk factor for the progression of lesions <50%S. Apo C-III and its distribution between HDL and LDL-VLDL is a marker of triglyceride-rich lipoprotein metabolism.4 Furthermore, apo C-III in VLDL is associated with smaller, more dense VLDL subclasses believed to be particularly atherogenic.22 23 In the present study, both small VLDL mass (Sf 20 to 60) and intermediate VLDL mass (Sf 60 to 100) were associated with progression of lesions <50%S. At baseline, the correlations between apo C-III in the LDL-VLDL fraction and VLDL subclasses were .72 for total VLDL mass, .68 for Sf 20 to 60 mass (small VLDL), .70 for Sf 60 to 100 mass (intermediate VLDL), and .67 for Sf 100 to 400 mass (large VLDL) (all P<.0001). Therefore, by two completely independent methods of identifying triglyceride-rich lipoproteins on the basis of different physiochemical properties, one based on protein composition (electroimmunoassay for apo C-III) and the other based on lipoprotein mass (analytical ultracentrifugation), the relationship between progression of lesions <50%S and triglyceride-rich lipoproteins is confirmed. In CLAS, apo C-III as well as non-HDL cholesterol were also associated with the progression of coronary artery lesions.24 It has recently been proposed that non-HDL cholesterol may provide a more meaningful method of risk assessment than the traditional measure of LDL cholesterol.25
Our previous report from MARS also indicated that both the total cholesterol/HDL cholesterol and LDL cholesterol/HDL cholesterol ratios were significantly associated with progression of both low-grade and high-grade coronary artery lesions.3 In the MARS sample, these ratios were moderately to strongly correlated with most of the lipoprotein subclasses (for the total cholesterol/HDL cholesterol ratio, absolute r=.11 to .74, median r=.56 in the placebo group and absolute r=.10 to .65, median r=.47 in the lovastatin group; for the LDL cholesterol/HDL cholesterol ratio, absolute r=.02 to .71, median r=.45 in the placebo group, and absolute r=.07 to 0.66, median r=.27 in the lovastatin group). These lipoprotein ratios thus appear to be excellent clinical indicators of the likelihood of disease progression by simultaneously reflecting the contribution of each individual lipoprotein subclass to the progression of CAD. However, these ratios do not elucidate which specific lipoproteins may be the most important contributors to progression of CAD. We therefore used our measurements of lipoprotein subclasses to show that both triglyceride-rich lipoprotein (specifically small VLDL) and HDL3 masses were the specific lipoproteins associated with progression of disease in the MARS cohort. Such an analysis of lipoprotein subclasses yields more specific hypotheses about pathophysiology and potential avenues of intervention than do analyses of lipoprotein ratios, which are not biologically meaningful parameters.
Although on-trial small LDL mass (Sf 0 to 3) was associated with CAD in this study, total LDL mass was not associated with progression of coronary artery lesions. Although LDL cholesterol is an established risk factor for CAD, it has failed to show a relation to coronary artery lesion progression in several studies.5 This has been particularly evident in studies that have been reported with separation of IDL from LDL.5 As in this study, IDL but not total LDL was associated with CAD progression; in our study this was evident in low-grade coronary artery lesions as well as in the all coronary artery lesion analysis. However, IDL and small LDL (LDL-IV) were significantly correlated in the MARS placebo group (r=.39 in the placebo group, P<.0001; r=.06 in the lovastatin group, NS). Associations between LDL levels and progression of CAD reported in some studies may be the result of increased levels of IDL cholesterol or of small dense LDL cholesterol included in the measurement of total LDL cholesterol. The lack of association between LDL and CAD progression also may be due in part to MARS subject selection in that eligibility criteria required all subjects to have elevated total cholesterol levels before randomization.
There is accumulating evidence that HDL3 has an equal or even stronger relationship to existing CAD (defined angiographically) than does HDL2.26 27 28 Our finding of an association between HDL3 and progression of coronary artery lesions is supported by epidemiological data29 30 31 and most recently by ancillary data from another serial coronary angiographic trial examining on-trial lipoprotein levels and the progression of CAD.19 In our study, HDL3 mass was the lipoprotein subclass least altered by lovastatin treatment relative to the placebo group (Table 2⇑). It is therefore possible that with reduction of all other lipoprotein subclasses, lovastatin therapy unmasked low HDL3 levels as a risk factor for progression of coronary artery lesions. While HDL2 masses were significantly correlated with VLDL subclasses in the MARS sample (r=−.20 to −.39 in the placebo group, .05<P<.0001; r=−.29 to −.38 in the lovastatin group, .01<P<.0001), HDL3 mass was not correlated to these triglyceride-rich lipoprotein masses (r=−.03 to −.23 in the placebo group, all but one correlation NS; r=.03 to .10 in the lovastatin group, all NS). This suggests that HDL3 mass reflects some aspect of lipoprotein metabolism that is likely independent of triglyceride-rich lipoproteins; HDL3 mass was thus, along with small VLDL mass, a statistically independent correlate of coronary artery lesion progression in the combined treatment group analysis.
Given the large number of associations tested in this study (14 lipoprotein subclass variables ×3 lesion size classes ×3 treatment group classes=126 associations), it is possible that the significant associations found might have occurred by chance. However, several points argue against this interpretation. First, at an α level of 0.05, there would be 6.3 (126×0.05) statistically significant findings expected by chance alone. Our data showed 22 statistically significant associations (Table 3⇑). Second, we performed multivariate stepwise analysis to identify the independently significant lipoprotein subclasses. Finally, a wealth of prior research, both from MARS3 and other studies,5 18 19 20 24 supports the above findings. Although guided by prior evidence, this was an exploratory study with no control for multiple hypothesis testing; these results should therefore be substantiated in future research.
The results of this study linked with a growing body of evidence3 5 18 19 20 24 indicate the importance of triglyceride-rich lipoproteins and HDL3 as independent risk factors for the progression of CAD. Additionally, these data indicate that more specific information concerning the pathogenesis of coronary artery lesion progression can be obtained from the study of specific apo B-containing lipoproteins.
Selected Abbreviations and Acronyms
|CAD||=||coronary artery disease|
|MARS||=||Monitored Atherosclerosis Regression Study|
|%S||=||percent diameter stenosis|
|QCA||=||quantitative coronary angiography|
|Sf||=||Svedberg flotation rate|
This study was supported in part by US Public Health Service grants NIH NHLBI-RO1-HL-49885, NIH NHLBI Program Project grant HL-18574, Merck Research Laboratories, and a grant from the National Dairy Promotion and Research Board and administered in cooperation with the National Dairy Council. This work was partially conducted at the Lawrence Berkeley National Laboratory through the US Department of Energy under Contract No. DE-AC03-76SF00098.
Cashin-Hemphill L, Kramsch DM, Azen SP, DeMets D, DeBoer LWV, Hwang I, Vailas L, Hirsch LJ, Mack WJ, Hodis HN, Mahrer PR, Selzer RH, Alaupovic P, Blankenhorn DH. The Monitored Atherosclerosis Regression Study (MARS): design, methods, and baseline results. The Online Journal of Current Clinical Trials. 1992: Doc 26.
Blankenhorn DH, Azen SP, Kramsch DM, Mack WJ, Cashin-Hemphill L, Hodis HN, DeBoer LWV, Mahrer PR, Masteller MJ, Vailas LI, Alaupovic P, Hirsch LJ, and the MARS Research Group. Coronary angiographic changes with lovastatin therapy: the Monitored Atherosclerosis Regression Study (MARS). Ann Intern Med. 1993;119:969-976.
Hodis HN, Mack WJ, Azen SP, Alaupovic P, Pogoda JM, LaBree L, Hemphill LC, Kramsch DM, Blankenhorn DH. Triglyceride- and cholesterol-rich lipoproteins have a differential effect on mild/moderate and severe lesion progression as assessed by quantitative coronary angiography in a controlled trial of lovastatin. Circulation. 1994;90:42-49.
Selzer RH, Hagerty C, Azen SP, Siebes M, Lee P, Shircore A, Blankenhorn DH. Precision and reproducibility of quantitative coronary angiography with applications to controlled clinical trials: a sampling study. J Clin Invest. 1989;83:520-526.
Lindgren FT, Jensen LC, Hatch FT. The isolation and quantitative analysis of serum lipoproteins. In: Nelson GJ, ed. Blood Lipids and Lipoproteins. New York, NY: John Wiley Interscience; 1972:181-274.
Brown BG, Zhao XQ, Sacco DE, Albers JJ. Lipid lowering and plaque regression: new insights into prevention of plaque disruption and clinical events in coronary disease. Circulation. 1993;87:1781-1791.
Azen SP, Mack WJ, Cashin-Hemphill L, LaBree L, Shircore AM, Selzer RH, Blankenhorn DH, Hodis HN. Progression of coronary artery disease predicts clinical coronary events: long-term follow-up from the Cholesterol Lowering Atherosclerosis Study (CLAS). Circulation. 1996;93:34-41.
Blankenhorn DH, Hodis HN. Arterial imaging and atherosclerosis reversal. Arterioscler Thromb. 1994;14:177-192.
Williams PT, Vranizan KM, Krauss RM. Correlations of plasma lipoproteins with LDL subfractions by particle size in men and women. J Lipid Res. 1992;33:765-774.
Mack WJ, Selzer RH, Pogoda JM, Lee PL, Shircore AM, Azen SP, Blankenhorn DH. Comparison of computer- and human-derived coronary angiographic end-point measures for controlled therapy trials. Arterioscler Thromb. 1992;12:348-356.
Kragel AH, Shanthasundari GR, Wittes JT, Roberts WC. Morphometric analysis of the composition of atherosclerotic plaques in the four major epicardial coronary arteries in acute myocardial infarction and in sudden coronary death. Circulation. 1989;80:1747-1756.
Tatami R, Mabuchi R, Ueda K, Udea R, Haba T, Kametani T, Ito S, Koizumi J, Ohta M, Miyamoto S, Nakayama A, Kanaya H, Oiwake H, Genda A, Takeda R. Intermediate-density lipoprotein and cholesterol-rich very low density lipoprotein in angiographically determined coronary artery disease. Circulation. 1981;64:1174-1184.
Reardon MF, Nestel PJ, Craig KH, Harper RW. Lipoprotein predictors of the severity of coronary artery disease in men and women. Circulation. 1985;71:881-888.
Steiner G, Schwartz L, Shumak S, Poapst M. The association of increased levels of intermediate-density lipoproteins with smoking and with coronary artery disease. Circulation. 1987;75:124-130.
Phillips NR, Waters D, Havel RJ. Plasma lipoproteins and progression of coronary artery disease evaluated by angiography and clinical events. Circulation. 1993;88:2762-2770.
Austin MA, King M-C, Vranizan KM, Krauss RM. Atherogenic lipoprotein phenotype: a proposed genetic marker for coronary heart disease risk. Circulation. 1990;82:495-506.
Le NA, Givson JC, Ginsberg HN. Independent regulation of plasma apolipoprotein C-II and C-III concentrations in very low density and high density lipoproteins: implications for the regulation of the catabolism of these lipoproteins. J Lipid Res. 1988;29:669-677.
Blankenhorn DH, Alaupovic P, Wickham E, Chin HP, Azen SP. Prediction of angiographic change in native human coronary arteries and aortocoronary bypass grafts: lipid and nonlipid factors. Circulation. 1990;81:470-476.
Brensike JF, Levy RI, Kelsey SF, Passamani ER, Richardson JM, Loh IK, Stone NJ, Aldrich RF, Battaglini JW, Moriarty DJ, Fisher MR, Friedman L, Friedewald W, Detre KM, Epstein SE. Effects of therapy with cholestyramine on progression of coronary arteriosclerosis: results of the NHLBI Type II Coronary Intervention Study. Circulation. 1984;69:313-324.
Kempen HJ, Van Gent CM, Buytenhek R, Buis B. Association of cholesterol concentrations in low-density lipoprotein, high-density lipoprotein, and high-density lipoprotein subfractions, and of apolipoproteins AI and AII, with coronary stenosis and left ventricular function. J Lab Clin Med. 1987;109:19-26.
Drexel H, Amann FW, Beran J, Rentsch K, Candina R, Muntwyler J, Luethy A, Gasser T, Follath F. Plasma triglycerides and three lipoprotein cholesterol fractions are independent predictors of the extent of coronary atherosclerosis. Circulation. 1994;90:2230-2235.
Sweetnam PM, Bolton CH, Yarnell JWG, Bainton D, Baker IA, Elwood PC, Miller NE. Associations of the HDL2 and HDL3 cholesterol subfractions with the development of ischemic heart disease in British men: the Caerphilly and Speedwell Collaborative Heart Disease Studies. Circulation. 1994;90:769-774.
Buring JE, O’Connor GT, Goldhaber SZ, Rosner B, Herbert PN, Blum CB, Breslow JL, Hennekens CH. Decreased HDL2 and HDL3 cholesterol, apo A-I and apo A-II and increased risk of myocardial infarction. Circulation. 1992;85:22-29.