Abstract
Objective
The significant cardiovascular disease (CVD) event reduction in VA-HIT could not be fully explained by the 6% increase in HDL-C with the fibrate, gemfibrozil. We examined whether measurement of HDL subpopulations provided additional information relative to CVD-risk reduction.
Methods and Results
HDL subpopulations were characterized by 2-dimensional gel-electrophoresis in subjects who were treated with gemfibrozil (n=754) or placebo (n=741). In this study, samples obtained at the 3-month visit were used and data were analyzed prospectively using CVD events (CHD death, MI, or stroke) during the 5.1 years follow up. Analyses in the gemfibrozil arm showed that subjects with recurrent CVD events had significantly higher preβ-1 and had significantly lower α-1 and α-2 HDL levels than those without such events. Preβ-1 level was a significant positive predictor; α-1 and α-2 levels were significant negative risk factors for future CVD events. α-2 level was superior to HDL-C level in CVD-risk assessment after adjustment for established risk factors. Gemfibrozil treatment was associated with 3%-6% decreases in the small, lipid-poor preβ-1 HDL and in the large, lipid-rich α-1 and α-2 HDL and with increases in the small α-3 (3%) and preα-3 (16%) HDLs.
Conclusions
While the use of gemfibrozil has been associated with reduction in CVD events in VA-HIT, HDL subpopulation analysis indicates that gemfibrozil-mediated improvement in CVD risk might not be the result of its effects on HDL. It is quite possible that much of the cardiovascular benefits of gemfibrozil are due to a much wider spectrum of effects on metabolic processes that is not reflected by changes in blood lipids and HDL subpopulations.
Keywords: gemfibrozil, HDL-C, HDL subpopulations, CVD risk
Introduction
VA-HIT demonstrated that gemfibrozil therapy significantly reduced the 5-year incidence of major coronary heart disease (CHD) events and produced a significant reduction in stroke in men with known CHD and low HDL-C and low-density lipoprotein cholesterol (LDL-C) levels (1). HDL-C was raised by 6% with gemfibrozil and this increase partially predicted the reduction in CHD events (2). Participants in VA-HIT had a high prevalence of diabetes and the features of the metabolic syndrome. VA-HIT provided an opportunity for investigating the relationship between HDL-related parameters and cardiovascular risk in a well-characterized cohort of CHD patients selected with low HDL-C and low LDL-C levels.
HDL is a heterogeneous class of lipoprotein particles with subspecies that differ in apolipoprotein and lipid composition, size, density, and charge. The different subspecies of HDL appear to have different physiologic functions (3-6). In the last decade several methods have been developed to identify specific HDL subspecies and explore their role in the etiology of cardiovascular disease (CVD) including quantitative non-denaturing 2-dimensional gel-electrophoresis and image analysis (7, 8). Apolipoprotein (apo) A-I-containing HDL subpopulations were assessed in a subset (n=1495) of VA-HIT by 2d gel-electrophoresis - which characterizes plasma HDL subpopulations by electrophoretic mobility (preβ-, α-, and preα-mobility) and size (4-17nm). ApoA-I content in the different particles was quantitatively determined by image-analysis. It was demonstrated that those subjects who developed new CVD events had higher concentrations of apoA-I in the poorly-lipidated, small α-3 and preβ-1 HDL particles and less apoA-I in the more-lipidated and larger α-1 and α-2 HDL particles as compared to subjects who did not develop such events (9). Indeed, for every 1 standard deviation (SD) increase in α-1 and HDL-C, the hazard for new CVD events decreased by 18% and 15%, respectively.
The present analysis was performed to assess the effects of gemfibrozil treatment on the apoA-I-containing HDL subpopulations in a subset of VA-HIT to determine whether the reduction in CVD events in this trial could be related to a change in the HDL subpopulation profile that was a consequence of gemfibrozil therapy.
Methods
Study Population
The VA-HIT study design and population have been described in detail (10). Briefly, men were recruited at 20 VA medical centers throughout the US. Eligibility for the trial required a documented history of CHD (including previous myocardial infarction [MI], coronary revascularization, or angiographic evidence of stenosis >50% of the luminal diameter in ≥1 major epicardial coronary arteries), an absence of serious coexisting conditions, an HDL-C level ≤40 mg/dL (1 mmol/L), an LDL-C level ≤140 mg/dL (3.6 mmol/L), a triglyceride (TG) level ≤300 mg/dL (3.4 mmol/L), and an age <74 years. Participants in VA-HIT were randomly assigned to 2 groups receiving either gemfibrozil or matching placebo treatment. In these analyses, all subjects with recurrent CVD events (CVD+) (CHD death, MI, or stroke) were included in both arms and subjects without recurrent events (CVD-) were randomly selected in numbers based on power calculations. HDL subpopulations were determined in 741 (n=230 CVD+) subjects in the placebo and in 754 (n=168 CVD+) subjects in the gemfibrozil arm. All measurements were performed in samples obtained at the 3-month visit and subjects were followed for an average of 5.1 years. We had a prospective study design; therefore the selection of an early time point (3-month) provided a long follow up and a better compliance.
Laboratory measurements
Total-C, TG, and HDL-C concentrations were determined by standard enzymatic methods. HDL-C was isolated from the supernatant after dextran-sulfate magnesium precipitation. LDL-C was calculated according to the Friedewald formula. Total plasma apoA-I concentrations were measured with a turbidimetric immunoassay (Wako Diagnostics, Richmond, VA) on a Hitachi 911 analyzer. ApoA-I containing HDL subpopulations were determined by 2d non-denaturing gel electrophoresis, immunoblotting, and image analysis as described (7, 11). ApoA-I levels in the individual HDL subpopulations were calculated by multiplying plasma apoA-I levels by the subpopulation percentiles. Since each HDL particle has a fixed number of apoA-I molecules, the change in apoA-I levels in each HDL subpopulation is proportional to changes of particle numbers. The inter- and intra-assay coefficients of variation were <4% for the lipid measurements and <10% for the apoA-I and HDL subpopulation determinations. All plasma samples were stored at −80°C and were never thawed until analysis. The effects of long-term storage on HDL subspecies showed no significant changes in the values obtained after measurements of the same samples fresh and after short-term and long-term storage (7).
Statistical Analysis
Descriptive statistics including means ± SD for continuous variables or proportions for categorical variables were computed for all study variables: 1) in subjects in the placebo arm versus subjects in the gemfibrozil arm; and 2) in subjects with and without new CVD events (CHD death, MI, or stroke) in the gemfibrozil arm. The distribution of the variables was compared using 2-sample t-tests for continuous variables and χ2 tests for categorical variables. Cox proportional hazard models were used to determine the hazard ratios (HRs) for new CVD events in follow-up (5.1 years) based on a 1 SD increase in lipid and HDL subpopulation variables in the gemfibrozil arm. These variables were divided into quartiles for further analysis with the upper quartile compared to the lower quartile in Cox models adjusted for CHD-risk factors (age, smoking, hypertension, body mass index [BMI], and diabetes). Cochran-Armitage trend tests were performed using the percentage of CVD events in each quartile. A 4-model approach was used to compare the risk of 1 SD increase in α-2 with HDL-C: in model 1, data were unadjusted; in model 2, data were adjusted for non-lipid CHD-risk factors (age, smoking, hypertension, BMI, and presence of diabetes); in model 3, data were adjusted for lipid (HDL-C, LDL-C, and TG) and non-lipid risk factors; and in model 4, data were further adjusted for either α-2 for HDL-C or HDL-C for α-2. Finally, a Receiver Operating Curve (ROC) analysis was performed, using HDL-C, TG, and HDL subpopulations (preβ-1, α-1, α-2, and α-3) as variables in the gemfibrozil and in the placebo arm of the study, with a threshold of >0.75 set for significance.
The SAS statistical package version 9.1 was used in all analyses. Results with P values <0.05 were considered statistically significant.
All measurements were conducted in a blinded fashion and data were analyzed at the VA Cooperative Studies Coordinating Center (West Haven, CT). The study was approved and continually monitored by the subcommittee on human studies at Tufts University/New England Medical Center. All subjects gave written informed consent.
Results
None of the parameters were different at baseline between the two arms as the gemfibrozil and placebo arms were evenly matched for clinical characteristics and laboratory values at baseline in VA-HIT (2). Therefore, we have compared the gemfibozil arm to the placebo arm at the 3-month visit to assess the influence of gemfibrozil on the measured variables in this sub-study (Table 1). The average age of the subjects was 64 years, about 35% of them had diabetes and about 60% had hypertension. Subjects who received gemfibrozil treatment had higher mean HDL-C (6% p<0.001) and lower mean TG (-33% p<0.001) levels than subjects who received placebo. Subjects receiving gemfibrozil had lower mean preβ-1 (-6% p<0.05), α-1 (4.5% ns), and α-2 (-3% p<0.05) and higher mean α-3 (3% p<0.05) and preα-3 (16% p<0.001) HDL subpopulation levels than those receiving placebo.
Table 1.
Characteristics of subjects after 3 months of treatment with placebo or gemfibrozil in VA-HIT.
| Variable | Placebo | Gemfibrozil |
|---|---|---|
| N=741 | N=754 | |
| Age (year) | 64.2 ± 7.0 | 64.0 ± 7.4 |
| BMI (kg/m2) | 29.4 ± 4.7 | 29.1 ± 5.0 |
| Hypertension (%) | 59.9 | 56.8 |
| Diabetes (%) | 35.0 | 34.5 |
| Smoking (%) | 17.8 | 19.1 |
| Total-C | 175.5 ± 27.0 | 166.8 ± 27.7* |
| TG† | 163.7±76.6 | 109.5 ± 58.7* |
| LDL-C | 111.6 ± 25.0 | 111.6 ± 25.2 |
| HDL Variables | ||
| HDL-C | 31.5 ± 5.7 | 33.3 ± 6.4* |
| Apolipoprotein A-I | 109.5 ± 17.6 | 109.6 ± 17.8 |
| Preβ-1 | 12.7 ± 6.6 | 12.0 ± 6.4** |
| Preβ-2 | 2.1 ± 1.3 | 2.2 ± 1.4 |
| α-1 | 9.0 ± 5.0 | 8.6 ± 4.8 |
| α-2 | 31.8 ± 8.3 | 30.8 ± 8.5** |
| α-3 | 40.3 ± 9.9 | 41.4 ± 9.8** |
| Preα-1 | 2.8 ± 2.3 | 2.9 ± 2.5 |
| Preα-2 | 5.2 ± 2.3 | 5.4 ± 2.4 |
| Preα-3 | 5.5 ± 2.2 | 6.4 ± 2.5* |
Values are mean ± SD (mg/dl) or as indicated.
Statistical test performed using log transformed values.
p<0.001
p<0.05
Conversion factor from mg/dl to mmol: for cholesterol, divide numbers by 38.88; for TG, divide numbers by 86.88.
Table 2 compares subjects with new CVD events to those without such events in the gemfibrozil arm of the study. Among the 754 subjects studied in the gemfibrozil arm, 168 subjects experienced an MI, stroke or CHD death in the 5.1-year follow-up. There were no significant differences in the measured lipid parameters between the 2 groups; however, there were significant differences in the mean concentrations of several HDL subpopulations: subjects with new CVD events had higher level of preβ-1 (14% p<0.01) and lower levels of preβ-2 (-13% p< 0.05), α-1 (-10% p<0.05), α-2 (-8% p<0.01), and preα-2 (-9% p<0.05) HDL subpopulations than those without new CVD events.
Table 2.
Characteristics of subjects in the gemfibrozil arm with and without new CVD events (non-fatal MI, CHD death and stroke) in the 5.1 year follow up.
| Variable | CVD (+) | CVD (-) |
|---|---|---|
| N=168 | N=586 | |
| Age (year) | 64.8 ± 7.0 | 63.7 ± 7.5 |
| BMI (kg/m2) | 29.0 ±5.0 | 29.1 ± 5.0 |
| Hypertension (%) | 63.0 | 54.9 |
| Diabetes (%) | 38.7 | 33.3 |
| Smoking (%) | 23.8 | 17.7 |
| Total-C | 169.2 ± 27.5 | 166.2 ± 27.8 |
| LDL-C | 113.0 ± 24.5 | 111.2 ± 25.4 |
| TG† | 119.2 ± 69.9 | 106.8 ± 54.8 |
| HDL Variables | ||
| HDL-C | 32.5 ± 6.0 | 33.6 ± 6.4 |
| Apolipoprotein A-I | 107.5 ± 16.9 | 110.2 ± 18.0 |
| Preβ-1 | 13.3 ± 7.0 | 11.6 ± 6.1* |
| Preβ-2 | 2.0 ± 1.3 | 2.3 ± 1.4** |
| α-1 | 7.9 ± 4.3 | 8.8 ± 4.9** |
| α-2 | 28.9 ± 8.2 | 31.4 ± 8.5* |
| α-3 | 41.7 ± 9.1 | 41.3 ± 10.0 |
| Preα-1 | 2.7 ± 2.3 | 2.9 ± 2.5 |
| Preα-2 | 5.0 ± 2.4 | 5.5 ± 2.4** |
| Preα-3 | 6.1 ± 2.4 | 6.5 ± 2.5 |
Values are mean ± SD (mg/dl) or as indicated.
Statistical test performed using log transformed values.
p<0.01
p<0.05
Cox proportional hazard models were used to determine the HRs for new CVD events in follow-up for 1 SD unit increase in the measured parameters in the gemfibrozil arm of the study (Table 3). These analyses indicated that HDL-C (HR=0.83 p<0.01), TG (HR=1.21 p<0.008), preβ-1 (HR=1.18 p<0.01), preβ-2 (HR=0.80 p<0.01), and α-2 (HR=0.81 p<0.01) were independent predictors of new CVD events after adjusting data for established non-lipid CHD-risk factors (age, smoking, hypertension, BMI, and diabetes). HDL-C lost power (p=0.23) to predict new CVD events when data were further adjusted for LDL-C and logTG (Table 4, model 3). α-2 HDL lost power (p<0.06) to predict new CVD events after data were adjusted for the above parameters as well as HDL-C (Table 4, model 4).
Table 3.
Hazard ratios for lipids, apoA-I, and HDL subpopulations in predicting cardiovascular endpoints (n=168) in the gemfibrozil arm (n=754)*.
| Unadjusted Analysis | Adjusted Analysis** | |||
|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | |
| Total-C | 1.09 (0.94,1.26) | 0.24 | 1.13 (0.95,1.33) | 0.13 |
| LDL-C | 1.05 (0.91,1.22) | 0.53 | 1.08 (0.93,1.26) | 0.33 |
| TG | 1.20 (1.05,1.38) | 0.008 | 1.21 (1.05,1.38) | 0.008 |
| HDL Variables | ||||
| HDL-C | 0.81 (0.70,0.93) | 0.003 | 0.83 (0.72,0.96) | 0.01 |
| ApoA-I | 0.88 (0.75,1.03) | 0.12 | 0.88 (0.75,1.03) | 0.12 |
| Preβ-1 | 1.19 (1.05,1.36) | 0.008 | 1.18 (1.04,1.35) | 0.01 |
| Preβ-2 | 0.81 (0.68,0.96) | 0.02 | 0.80 (0.67,0.95) | 0.01 |
| α-1 | 0.85 (0.72,1.01) | 0.06 | 0.86 (0.72,1.01) | 0.07 |
| α-2 | 0.79 (0.68,0.93) | 0.004 | 0.81 (0.69,0.96) | 0.01 |
| α-3 | 1.02 (0.88,1.19) | 0.79 | 0.99 (0.86,1.16) | 0.95 |
| Preα-1 | 0.93 (0.79,1.09) | 0.38 | 0.93 (0.79,1.10) | 0.42 |
| Preα-2 | 0.85 (0.72,0.99) | 0.05 | 0.88 (0.75,1.04) | 0.13 |
| Preα-3 | 0.88 (0.75,1.03) | 0.11 | 0.88 (0.75,1.04) | 0.12 |
Hazard ratios (HR) were estimated using Cox proportional hazard models. Cardiovascular endpoints were defined as stroke, non-fatal MI or CHD death. 95% Confidence Interval (CI) is given 1 SD units for each variable.
Data were adjusted for age, smoking, hypertension, BMI, diabetes.
Table 4.
Hazard ratios as calculated for α-2 HDL and HDL-C in predicting cardiovascular endpoints (n=168) in the gemfibrozil arm (n=754)*.
| Model** | HR for each 1SD increase in α-2 (8.51) | HR for each 1SD increase in HDL-C (6.37) | ||
|---|---|---|---|---|
| HR (95% CI) | p-value | HR (95% CI) | p-value | |
| Model-1 | 0.79 (0.68,0.93) | .004 | 0.81 (0.70,0.93) | .003 |
| Model-2 | 0.81 (0.69,0.96) | .01 | 0.83 (0.72,0.96) | .01 |
| Model-3 | 0.83 (0.71,0.98) | .03 | 0.90 (0.76,1.07) | .23 |
| Model-4 | 0.82 (0.66,1.01) | .06 | 0.93 (0.76,1.14) | .50 |
Hazard ratios (HR) were estimated using Cox proportional hazard models. Cardiovascular endpoints were defined as stroke, non-fatal MI or CHD death. Confidence Interval (CI) is given 1 SD units in α-2 (8.51) and HDL-C (6.37).
Model-1: Data were unadjusted; Model-2: Data were adjusted for non-lipid CHD-risk factors (age, smoking, hypertension, BMI, and diabetes); Model-3: Data were further adjusted for LDL-C and logTG; Model-4: For α-2 data were further adjusted for HDL-C and for HDL-C data were further adjusted for α-2.
The relationship between HDL subpopulation quartiles and CVD event rate is shown in Table 5. Subjects with the lowest preβ-1 level had 47% lower relative risk (RR) for recurrent CVD events than subjects with the highest preβ-1 level (RR=1.95 p=0.005). In contrast, subjects with the highest levels of preβ-2 (57%, RR=0.55 p=0.02), α-2 (83%, RR=0.57 p=0.002), and preα-2 (70%, RR=0.58 p=0.02) had lower RR than subjects with the lowest levels of these parameters. The Cochran-Armitage trend test indicated a positive association between recurrent CVD events and preβ-1 (p=0.003) and an inverse trend for preβ-2 (p=0.04), α-1 (p=0.05), α-2 (p=0.0002), and preα-2 (p=0.03).
Table 5.
Quartile analysis for evaluating the relative risk (RR) for recurrent CVD events (%) in the gemfibrozil arm of VA-HIT.
| Quartile-1 | Quartile-2 | Quartile-3 | Quartile-4 | RR (95% CI) | p-value* | Trend-Test ** P= | |
|---|---|---|---|---|---|---|---|
| Percent (%) new CVD events | |||||||
| Preβ-1† | 14.9 | 22.2 | 24.1 | 27.9 | 1.95 (1.22,3.11) | 0.005 | 0.003 |
| Preβ-2 | 23.4 | 26.6 | 24.2 | 14.9 | 0.55 (0.33,0.89) | 0.02 | 0.04 |
| α-1 | 24.9 | 25.5 | 21.3 | 17.5 | 0.70 (0.44,1.11) | 0.13 | 0.05 |
| α-2 | 29.1 | 27.7 | 16.5 | 15.9 | 0.57 (0.36,0.90) | 0.02 | 0.0002 |
| α-3 | 19.1 | 23.4 | 24.5 | 22.2 | 1.03 (0.66,1.63) | 0.89 | 0.43 |
| preα-1 | 24.9 | 20.7 | 26.5 | 17.0 | 0.69 (0.37,1.09) | 0.11 | 0.19 |
| preα-2 | 28.0 | 20.0 | 24.6 | 16.5 | 0.58 (0.37,0.92) | 0.02 | 0.03 |
| preα-3 | 23.8 | 27.7 | 18.0 | 19.7 | 0.80 (0.52,1.25) | 0.33 | 0.10 |
N=754 including 168 subjects with recurrent CVD events (stroke, non-fatal MI or CHD death).
P-values for RR were calculated by comparing data in quartile 4 to data in quartile 1. Data were adjusted for age, smoking, hypertension, BMI, and diabetes.
Cochran-Armitage trend test.
The trend between particle concentrations and recurrent CVD events is positive.
ROC curve analysis using HDL-C, TG, and the major HDL subpopulations (preβ-1, α-1, α-2, and α-3) as variables did not distinguish between subjects with and without new CVD events either in the placebo or in the gemfibrozil arm. None of the c-values (area under the curve) were higher than 0.580, while 0.75 is the minimum for a positive threshold effect (data not shown).
Discussion
VA-HIT was the first lipid intervention trial to test whether increasing HDL-C concentrations in men selected with established CHD, low LDL-C and low HDL-C levels decrease CVD risk (2). The VA-HIT investigators concluded that the gemfibrozil-mediated reduction (22%) in new coronary events was partly dependent on an HDL-C increase since the benefit was independent of changes in the concentration of TGs or LDL-C as well as other major risk factors (1). In contrast to VA-HIT, in the Fenofibrate Intervention in Event Lowering in Diabetes (FIELD) trial, only a non-significant 11% reduction in coronary events was observed (12). As pointed out in a recent review, there have been substantial differences in results of fibrate trials and all fibrates may not have equivalent clinical benefit (13).
We examined whether measurement of HDL subpopulations by 2d gel-electrophoresis provided additional information relative to CVD-risk reduction by gemfibrozil in the VA-HIT. Gemfibrozil treatment was associated with 3%-6% decreases in the small, lipid-poor preβ-1 HDL and in the large, lipid-rich α-1 and α-2 HDL and with increases in the small α-3 (3%) and preα-3 (16%) HDLs.
Data generated in this study are in agreement with previous assessments of HDL subfractions: increases in HDL3-C but not in HDL2-C estimated after separation by differential polyanion precipitation, were significantly related to the development of new CVD events in VA-HIT (2). In a subset of VA-HIT, HDL particle number was assessed by nuclear magnetic resonance (NMR), which indicated a 10% increase in total HDL particle number and a 21% increase in the number of the small HDL subclasses in the gemfibrozil arm compared to the placebo arm (14). HDL subpopulation analysis by 2d gel-electrophoresis revealed differences in the effects of gemfibrozil on the small, lipid-poor HDL particles, which can not be differentiated by polyanion precipitation or by NMR. Among the three varieties of small HDL particles (preβ-1, α-3, and preα-3), only preβ-1 concentration was significantly lower while the concentrations of α-3 and preα-3 were significantly higher in the gemfibrozil arm compared to the placebo arm. Data generated by polyanion precipitation and NMR can not be directly compared to data generated by 2d gel-electrophoresis since the former methods measure lipid content and the latter one measures apoA-I in HDL. However, we have compared ultracentifugally separated HDL subclasses (HDL2 and HDL3) with to 2d-gel electrophoresis (7). We have shown that HDL3 was a composite of α-2- and the small α-3- and preβ-1-sized particles; HDL2 was composed mainly of the large α-1 and preα-1 particles.
Preβ-1 promotes cellular phospholipids and cholesterol efflux via the ATP-binding cassette transporter A1 (ABCA1) metabolic pathway and these additional lipids transform preβ-1 (mean diameter 5.4 nm) into more-lipidated α-3 HDL particles (mean diameter 6.7 nm) (15). The fractional catabolic rate (FCR) of apoA-I is inversely correlated with HDL particle size (16), thus transformation of preβ-1 into α-3 increases HDL residence time. In case-control studies, high preβ-1 level is associated with prevalent CHD (17). High preβ-1 level was a significant predictor for recurrent CVD events in subjects in the placebo and gemfibrozil arms combined (9) as well as in the gemfibrozil arm alone. We hypothesize that the gemfibrozil-mediated transformation of preβ-1 into α-3 is beneficial if the newly formed α-3 can further maturate into more-lipidated α-2 and α-1 HDL particles, which promote selective cholesterol uptake in the liver via the scavenger receptor class B type 1 (SRB1) pathway (15, 18). However, in this study, concentrations of the large, cholesterol-rich particles (α-1 and α-2) were lower in the gemfibrozil arm than in the placebo arm indicating a defect in full maturation of HDL particles. Previously, we have reported that a low α-1 level was a significant predictor for first CHD events in male participant of the Framingham Offspring Study (FOS) and a low level α-2 was a significant predictor for recurrent CVD events in subjects in the combined arms of VA-HIT (8, 9).
Using NMR, Otvos et al. have reported that gemfibrozil-mediated increase in small HDL particle number was significantly associated with decreased CVD events in a subgroup of VA-HIT (14). In our assessment, only α-3 and preα-3 concentrations increased among small HDL particles which particles have been shown to be positively associated with CVD risk (7-9). These seemingly conflicting results make it difficult to explain the beneficial effects of gemfibrozil on CVD risk by its effects on HDL. Moreover, a ROC curve analysis indicated that neither HDL-C, TG, nor the major HDL subpopulation levels distinguished between subjects with new events and subjects without new events since none of the c-values reached the minimum of 75% for positive threshold effect.
Our findings suggest that gemfibrozil has a significantly different effect on the HDL-subpopulation profile than do other lipid-modifying drugs we have investigated (statins and nicotinic acid) (19-21). Statins and nicotinic acid significantly decrease preβ-1 levels and significantly increase α-1 and α-2 HDL levels in subjects with CHD resulting in a shift in the HDL subpopulation profile towards normal. In the HDL-Atherosclerosis Treatment Study (HATS), we have documented that the increase in α-1 level were significantly correlated with decrease in coronary artery stenosis (19). We (20, 21) and others (22) have similarly found that statins increase HDL size, or more specifically α-1 concentrations, and suggested that this might be a consequence of a reduction in TG concentrations that could, in turn, decreases of cholesteryl ester transfer protein (CETP) activity (22). Data on subjects treated with a specific CETP inhibitor, clearly support the above statement (23). In contrast, gemfibrozil treatment was associated with decreases in the large (α-1 and α-2) and increases in the small, lipid-poor HDL particles (α-3, preα-3) despite of a 33% reduction in TG, suggesting that a decrease in TG levels is not necessarily accompanied by a decrease in CETP activity and an increase in the concentration of large HDL particles. This assumption is supported by reports on fibrate mechanism indicating that gemfibrozil did not significantly decreased CETP (24).
Based on this and previous studies, we believe that several HDL subpopulations are involved in the development of CVD but probably via different mechanisms. HDL particles may possess several potentially athero-protective properties, not necessarily distributed evenly among the different HDL particles. The most important and experimentally verified athero-protective functions of HDL are: 1) mediating reverse cholesterol transport, 2) inhibiting LDL oxidation, 3) improving endothelial function, and 4) decreasing inflammation in the vessel wall. More studies are needed to elucidate the effects of gemfibrozil on other possibly beneficial changes in HDL function, as fibrates have been shown to act as anti-inflammatory agents by inhibiting the NFκB inflammatory cascade (25), which is the master regulator of production of several pro-inflammatory proteins. PPAR-α -which is upregulated by gemfibrozil- binds to the P65 unit of NFκB and inhibits the translocation of NFκB into the nucleus; therefore NFκB can not activate genes of pro-inflammatory proteins (26). Moreover, fibrates improve coagulation and fibrinolysis (27, 28), increase LDL size in diabetic patients (29), and increase insulin sensitivity by decreasing free fatty acid production by decreasing NF-κB-mediated lipoprotein lipase production.
These data confirm that apoA-I-containing HDL subpopulations are related to CVD risk, but do not support earlier analyses of the same subjects that HDL played a significant role in the gemfibrozil-mediated CVD-risk reduction. It is quite possible that much of the cardiovascular benefits of gemfibrozil are due to the considerable decrease in small dense-LDL (14) concentrations, or a much wider spectrum of effects on metabolic processes that is not reflected by changes in blood lipids and HDL subpopulations, such as decreases in inflammation.
Acknowledgments
This study was supported by the National Institutes of Health/National Heart, Lung, and Blood Institute (HL-64738 PI: Asztalos) and by the VA Cooperative Studies Program of the Department of Veterans Affairs Office of Research and Development, Washington, DC.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Rubins HB, Robins SJ, Collins D, Fye CL, Anderson JW, Elam MB, Faas FH, Linares E, Schaefer EJ, Schectman G, Wilt TJ, Wittes J. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of high-density lipoprotein cholesterol. N Eng J Med. 1999;341:410–418. doi: 10.1056/NEJM199908053410604. [DOI] [PubMed] [Google Scholar]
- 2.Robins SJ, Collins D, Wittes JT, Papademetriou V, Deedwania PC, Schaefer EJ, McNamara JR, Kashyap ML, Hershman JM, Wexler LF, Rubins HB for the VA-HIT Study Group. Relation of gemfibrozil treatment and lipid levels with major coronary events. VA-HIT: A randomized controlled trial. JAMA. 2001;285:1585–1591. doi: 10.1001/jama.285.12.1585. [DOI] [PubMed] [Google Scholar]
- 3.Mowri HO, Patsch W, Smith LC, Gotto AM, Patsch JR. Different reactivities of high-density lipoprotein2 subfractions with hepatic lipase. J Lipid Res. 1992;33:1269–1279. [PubMed] [Google Scholar]
- 4.Castro GR, Fielding CJ. Early incorporation of cell-derived cholesterol into preβ-migrating high density lipoprotein. Biochemistry. 1988;27:25–29. doi: 10.1021/bi00401a005. [DOI] [PubMed] [Google Scholar]
- 5.Miida T, Kawano M, Fielding PE, Fielding CJ. Regulation of the concentration of preβ high-density lipoprotein in normal plasma by cell membranes and lecithin-cholesterol acyltransferase activity. Biochemistry. 1992;31:11112–11117. doi: 10.1021/bi00160a022. [DOI] [PubMed] [Google Scholar]
- 6.von Eckardstein A, Huang Y, Assmann G. Physiological role and clinical relevance of high-density lipoprotein subclasses. Curr Op Lipidol. 1994;5:404–416. doi: 10.1097/00041433-199412000-00003. [DOI] [PubMed] [Google Scholar]
- 7.Asztalos BF, Roheim PS, Milani RL, Lefevre M, McNamara JR, Horvath KV, Schaefer EJ. Distribution of apoA-I-containing HDL subpopulations in patients with coronary heart disease. Arterioscler Thromb Vasc Biol. 2000;20:2670–2676. doi: 10.1161/01.atv.20.12.2670. [DOI] [PubMed] [Google Scholar]
- 8.Asztalos BF, Cupples LA, Demissie S, Horvath KV, Cox CE, Batista MC, Schaefer EJ. High-density lipoprotein subpopulation profile and coronary heart disease prevalence in male participants in the Framingham Offspring Study. Arterioscler Thromb Vasc Biol. 2004;24:2181–2187. doi: 10.1161/01.ATV.0000146325.93749.a8. [DOI] [PubMed] [Google Scholar]
- 9.Asztalos BF, Collins D, Cupples LA, Demissie S, Horvath KV, Bloomfield HE, Robins SJ, Schaefer EJ. Value of high density lipoprotein (HDL) subpopulations in predicting recurrent cardiovascular events in the Veterans Affairs HDL Intervention Trial. Arterioscler Thromb Vasc Biol. 2005;25:2185–2191. doi: 10.1161/01.ATV.0000183727.90611.4f. [DOI] [PubMed] [Google Scholar]
- 10.Rubins HB, Robins SJ, Iwane MK, Boden WE, Elam MB, Fye CL, Gordon DJ, Schaefer EJ, Schectman G, Wittes JT. Rationale and design of the Department of Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial (HIT) for secondary prevention of coronary artery disease in men with low high-density lipoprotein cholesterol and desirable low-density lipoprotein cholesterol. Am J Cardiol. 1993;71(1):45–52. doi: 10.1016/0002-9149(93)90708-k. [DOI] [PubMed] [Google Scholar]
- 11.Asztalos BF, Sloop CH, Wong L, Roheim PS. Two-dimensional electrophoresis of plasma lipoproteins: recognition of new apoA-I-containing subpopulations. Biochim Biophys Acta. 1993;1169:291–300. doi: 10.1016/0005-2760(93)90253-6. [DOI] [PubMed] [Google Scholar]
- 12.Keech A, Simes RJ, Barter P, Best J, Scott R, Taskinen MR, Forder P, Pillai A, Davis T, Glasziou P, Drury P, Kesaniemi YA, Sullivan D, Hunt D, Colman P, d’Emden M, Whiting M, Ehnholm C, Laakso M FIELD study investigators. Effects of long-term fenofibrate therapy on cardiovascular events in 9795 people with type 2 diabetes mellitus (the FIELD study): randomised controlled trial. Lancet. 2005;366(9500):1849–1861. doi: 10.1016/S0140-6736(05)67667-2. [DOI] [PubMed] [Google Scholar]
- 13.Robins SJ, Bloomfield HE. Fibric acid derivatives in cardiovascular disease prevention: results from the large clinical trials. Curr Opin Lipidol. 2006;17:431–439. doi: 10.1097/01.mol.0000236370.27508.9d. [DOI] [PubMed] [Google Scholar]
- 14.Otvos JD, Collins D, Freedman DS, Shalaurova I, Schaefer EJ, McNamara JR, Bloomfield HE, Robins SJ. 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(12):1556–1563. doi: 10.1161/CIRCULATIONAHA.105.565135. [DOI] [PubMed] [Google Scholar]
- 15.Asztalos BF, de la Llera-Moya M, Dallal GE, Horvath KV, Schaefer EJ, Rothblat GH. Differential effects of HDL subpopulations on cellular ABCA1- and SR-BI-mediated cholesterol efflux. J Lipid Res. 2005;46(10):2246–2253. doi: 10.1194/jlr.M500187-JLR200. [DOI] [PubMed] [Google Scholar]
- 16.Brinton EA, Eisenberg S, Breslow JL. Human HDL cholesterol levels are determined by apoA-I fractional catabolic rate, which correlates inversely with estimates of HDL particle size. Effects of gender, hepatic and lipoprotein lipases, triglyceride and insulin levels, and body fat distribution. Arterioscler Thromb. 1994;(5):707–720. doi: 10.1161/01.atv.14.5.707. [DOI] [PubMed] [Google Scholar]
- 17.Miida T, Nakamura Y, Inano K, Matsuto T, Yamaguchi T, Tsuda T, Okada M. Pre beta 1-high-density lipoprotein increases in coronary artery disease. Clin Chem. 1996;42(12):1992–2005. [PubMed] [Google Scholar]
- 18.Schaefer EJ, Asztalos BF. Cholesteryl ester transfer protein inhibition, high-density lipoprotein metabolism and heart disease risk reduction. Curr Opin Lipidol. 2006;17:394–398. doi: 10.1097/01.mol.0000236364.63840.d8. [DOI] [PubMed] [Google Scholar]
- 19.Asztalos BF, Batista M, Horvath KV, Cox CE, Dallal GE, Morse JS, Brown GB, Schaefer EJ. Change in α-1 HDL concentration predicts progression in coronary artery stenosis. Arterioscler Thromb Vasc Biol. 2003;23:847–852. doi: 10.1161/01.ATV.0000066133.32063.BB. [DOI] [PubMed] [Google Scholar]
- 20.Asztalos BF, Horvath KV, McNamara JR, Roheim PS, Rubinstein JJ, Schaefer EJ. Effects of atorvastatin on the HDL subpopulation profile of coronary heart disease patients. J Lipid Res. 2002;43:1701–1707. doi: 10.1194/jlr.m200037-jlr200. [DOI] [PubMed] [Google Scholar]
- 21.Asztalos BF, Horvath KV, McNamara JR, Roheim PS, Rubinstein JJ, Schaefer EJ. Comparing the effects of five different statins on the HDL subpopulation profiles of coronary heart disease patients. Atherosclerosis. 2002;164:361–369. doi: 10.1016/s0021-9150(02)00149-1. [DOI] [PubMed] [Google Scholar]
- 22.Guerin M, Lassel TS, Le Goff W, Farnier M, Chapman MJ. Action of atorvastatin in combined hyperlipidemia: preferential reduction of cholesteryl ester transfer from HDL to VLDL1 particles. Arterioscler Thromb Vasc Biol. 2000;20:189–197. doi: 10.1161/01.atv.20.1.189. [DOI] [PubMed] [Google Scholar]
- 23.Brousseau ME, Schaefer EJ, Wolfe ML, Bloedon LT, Digenio AG, Clark RW, Mancuso JP, Rader DJ. Effects of an inhibitor of cholesteryl ester transfer protein on HDL cholesterol. N Engl J Med. 2004;350(15):1505–1515. doi: 10.1056/NEJMoa031766. [DOI] [PubMed] [Google Scholar]
- 24.Guerin M, Bruckert E, Dolphin PJ, Turpin G, Chapman MJ. Fenofibrate reduces plasma cholesteryl ester transfer from HDL to VLDL and normalizes the atherogenic, dense LDL profile in combined hyperlipidemia. Arterioscler Thromb Vasc Biol. 1996;16(6):763–772. doi: 10.1161/01.atv.16.6.763. [DOI] [PubMed] [Google Scholar]
- 25.Durrington PN, Mackness MI, Bhatnagar D, Julier K, Prais H, Arrol S, Morgan J, Wood GN. Effects of two different fibric acid derivatives on lipoproteins, cholesteryl ester transfer, fibrinogen, plasminogen activator inhibitor and paraoxonase activity in type IIb hyperlipoproteinaemia. Atherosclerosis. 1998;138(1):217–225. doi: 10.1016/s0021-9150(98)00003-3. [DOI] [PubMed] [Google Scholar]
- 26.Paumelle R, Blanquart C, Briand O, Barbier O, Duhem C, Woerly G, Percevault F, Fruchart JC, Dombrowicz D, Glineur C, Staels B. Acute antiinflammatory properties of statins involve peroxisome proliferator-activated receptor-alpha via inhibition of the protein kinase C signaling pathway. Circ Res. 2006;98(3):361–369. doi: 10.1161/01.RES.0000202706.70992.95. [DOI] [PubMed] [Google Scholar]
- 27.Schonfeld G. The effects of fibrates on lipoprotein and hemostatic coronary risk factors. Atherosclerosis. 1994;111(2):161–174. doi: 10.1016/0021-9150(94)90090-6. [DOI] [PubMed] [Google Scholar]
- 28.Turpin G, Bruckert E. Efficacy and safety of ciprofibrate in hyperlipoproteinaemias. Atherosclerosis. 1996;124:S83–87. doi: 10.1016/0021-9150(96)05861-3. [DOI] [PubMed] [Google Scholar]
- 29.Vakkilainen J, Steiner G, Ansquer JC, Aubin F, Rattier S, Foucher C, Hamsten A, Taskinen MR DAIS Group. Relationships between low-density lipoprotein particle size, plasma lipoproteins, and progression of coronary artery disease: the Diabetes Atherosclerosis Intervention Study (DAIS) Circulation. 2003;107:1733–1737. doi: 10.1161/01.CIR.0000057982.50167.6E. [DOI] [PubMed] [Google Scholar]