, Volume 48, Issue 2, pp 127–137

Statin Treatment Improves Plasma Lipid Levels but not HDL Subclass Distribution in Patients Undergoing Percutaneous Coronary Intervention


  • Li Tian
    • Laboratory of Endocrinology and Metabolism, West China HospitalSichuan University
    • State Key Laboratory of BiotherapySichuan University
  • Yucheng Chen
    • Cardiovascular Department, West China HospitalSichuan University
  • Chuanwei Li
    • Cardiovascular Department, West China HospitalSichuan University
  • Zhi Zeng
    • Cardiovascular Department, West China HospitalSichuan University
  • Yanhua Xu
    • Chengdu Hoist Biotechnology Co., Ltd.
  • Shiyin Long
    • Department of Biochemistry and Molecular BiologyUniversity of South China
    • Laboratory of Endocrinology and Metabolism, West China HospitalSichuan University
    • State Key Laboratory of BiotherapySichuan University
Original Article

DOI: 10.1007/s11745-012-3750-6

Cite this article as:
Tian, L., Chen, Y., Li, C. et al. Lipids (2013) 48: 127. doi:10.1007/s11745-012-3750-6


Despite the established efficacy of statin therapy, the risk of cardiovascular events remains high in many patients. We examined high-density lipoprotein (HDL) subclass distribution profiles among statin-treated coronary heart disease (CHD) patients undergoing percutaneous coronary intervention (PCI). Plasma HDL subclasses were measured in 85 patients with established CHD and quantified by two-dimensional gel electrophoresis and immunoblotting. In CHD patients with statin treatment, the mean value of total cholesterol (TC) reached the desirable level and the triacylglycerol level (TAG) was borderline high. Moreover, low density lipoprotein cholesterol (LDL-C), high density lipoprotein cholesterol (HDL-C), apolipoproteinA-I, and apolipoproteinB-100 levels in these patients resembled those in normolipidemic healthy subjects. The HDL subclass did not show a normal distribution and was characterized by the lower large-sized HDL2b contents and higher contents of small-sized preβ1-HDL in CHD patients, compared to those in normolipidemic control subjects. Multiple stepwise regression analysis revealed that the severity of coronary stenosis, determined by the Gensini Score, was significantly and independently predicted by HDL2b and HDL3b. Statin therapy was effective in modifying plasma lipids levels, but not adequate as a monotherapy to normalize the HDL subclass distribution phenotype of patients with CHD undergoing PCI. The HDL subclass distribution may aid in risk stratification, especially in patients with CHD and therapeutic LDL-C and HDL-C levels.


Coronary heart diseaseHigh-density lipoprotein subclassesStatins therapyQuantitative coronary angiographyTwo-dimensional gel electrophoresis-immunodetection



High density lipoprotein


Coronary heart disease


Percutaneous coronary intervention


Quantitative coronary angiography


Total cholesterol




Low density lipoprotein cholesterol


High density lipoprotein cholesterol






Acute myocardial infarction


Stable angina pectoris


Unstable angina pectoris




Low density lipoprotein(s)


High-density lipoprotein(s)


Reverse cholesterol transport


3-Hydroxy3-methylglutaryl coenzyme A


American College of Cardiology


American Heart Association


Thyroid stimulating hormone


National Cholesterol Education Program-Adult Treatment Panel III


Body mass index




Bovine serum albumin


Horseradish peroxidase


Immunoglobulin G


Standardized regression coefficient


ATP binding cassette transporter A1


Very low density lipoprotein


Standard deviation


The clinical benefits of cholesterol reduction have been established in large-scale primary and secondary intervention trials showing that treatment with cholesterol-lowering statin drugs results in decreased morbidity and mortality related to coronary heart disease (CHD) [1]. Statins are competitive inhibitors of 3-hydroxy3-methylglutaryl coenzyme A (HMG-CoA) reductase, a rate-limiting enzyme in cholesterol biosynthesis [2]. In large landmark trials, reducing low density lipoprotein (LDL) levels with statins decreased the incidence of major cardiovascular events by 25–45 % [35]. Nonetheless, even with an aggressive statin treatment regimen, considerable residual cardiovascular risks remain, including a high frequency of recurrent events [69]. Epidemiological evidence strongly favors the notion that the risk of cardiovascular disease (CVD) is inversely related to the plasma high-density lipoprotein (HDL) cholesterol concentration [10]. Low HDL cholesterol is still predictive of high CVD risk in subjects with low LDL cholesterol [11], as well as during statin treatment [12].

Circulating HDL particles are very heterogeneous with a very complex metabolic profile. The various HDL subclasses vary in quantitative and qualitative content of lipids, apolipoproteins, enzymes, and lipid transfer proteins, resulting in differences in shape, density, size, charge, and antigenicity. The variety of distinct particle subclasses may explain the multiple biological activities of HDL [13]. Although the compositional and functional heterogeneity of HDL particles is well known, HDL are often regarded as a single entity whose plasma levels can be reflected by a single measurement (HDL). Indeed, many pharmacological strategies that raise HDL have been evaluated without regard either to the functional specificity or to the heterogeneity of HDL particles [14]. However, evidence is accumulating supporting the concept that high HDL cholesterol levels do not always predict reduced CVD risk. The incremental decrease in end points through aggressive lipid lowering (IDEAL) trial and the European Prospective Investigation into Cancer and Nutrition (EPIC)-Norfolk case-control study revealed that (recurrent) CVD risk is not decreased in subjects with the highest HDL cholesterol and the greatest mean HDL particle size [15]. More recently, a high HDL cholesterol subgroup of individuals at increased risk for a first cardiovascular event was identified in the community-dwelling Prevention of Renal and Vascular End-Stage Disease (PREVEND) cohort using the “outcome event mapping approach,” a graphical exploratory data analysis tool that was originally developed by Corsetti et al. [16].

In this study, we focused on the frequency distribution of HDL particle subclasses in statin-treated CHD patients undergoing percutaneous coronary intervention (PCI). Our study suggested that monitoring the HDL subclass distribution might allow for a more accurate evaluation of cardiovascular risk in both CHD patients using statins and the premorbid population.


Study Design

The study was designed to investigate the HDL subclass distribution profile in a group of statin-treated CHD patients who underwent PCI and to compare the profile in CHD patients and healthy control subjects. CHD patients were further divided into unstable angina pectoris (UAP), stable angina pectoris (SAP), and acute myocardial infarction (AMI) subgroups according to the American College of Cardiology (ACC)/American Heart Association (AHA) Guidelines [17].

This was a single center, case-control study. The physicians and the patients were not blinded for treatment, but the laboratory staff carrying out the analyses received only number coded plasma samples. Protocols were reviewed and approved by the Chinese Ethics Committee of Registering Clinical Trials; and patients provided written informed consent prior to any study-related procedure.


Eighty-five patients with established CHD undergoing PCI participated in this study. These patients were hospitalized in the cardiovascular department of the West China Hospital, Sichuan University. Diagnosis was based on clinical history and was confirmed by quantitative coronary angiography (QCA). The results of angiographic examination were regarded positive for coronary atherosclerosis only if one or more major coronary arteries (right coronary, left main coronary, left anterior descending, and circumflex) had at least 50 % stenosis of the luminal area. The angiogram was reviewed by two physicians to obtain the Gensini score, which assesses the severity of coronary artery disease by grading the narrowing of the lumen of the coronary artery and scoring it as follows: 1 for 1–25 % narrowing, 2 for 26–50 % narrowing, 4 for 51–75 %, 8 for 76–90 %, 16 for 91–99 % and 32 for a completely occluded artery. This score was then multiplied by a factor according to the importance of the coronary artery as follows: 5 for a left main stem lesion, 2.5 for proximal left anterior descending artery and proximal circumflex artery lesions, 1.5 for a mid-left anterior descending artery lesion, and 1 for distal left anterior descending artery, mid/distal circumflex artery and right coronary artery lesions. The multiplication factor for any other branch was 0.5. Female subjects were only included in the study if they were postmenopausal or surgically sterile. Exclusion criteria included: uncontrolled diabetes mellitus, diabetes mellitus requiring insulin therapy, secondary causes of hyperlipidemia, such as uncontrolled primary hypothyroidism with thyroid stimulating hormone (TSH) greater than 5.5 uIU/ml; impaired renal function (creatinine >2.0 mg/dL) or nephritic syndrome; the presence of active liver disease or hepatic dysfunction; clinically significant hematology abnormalities. Additional exclusion criteria were uncontrolled hypertension (defined as a blood pressure of at least 140/90 mmHg), a current or recent history consumption of more than 14 alcoholic drinks per week, and use of immunosuppressive agents. All patients were concurrently taking antihypertensive medication (calcium channel blockers, 80 %; beta blockers, 85 %; diuretics, 75 %); 70 % were also using other antihypertensive therapies. Concurrent medications did not change over the course of the study.

Eighty age-, sex-matched normolipidemic subjects from West China Center of Medical Science, Sichuan University were recruited as the healthy control subjects. Waist, hip circumferences and the Gensini Score were not measured for these subjects. These subjects were not hyperlipidemic or diabetic, had no family history of premature CHD, and were not receiving lipid-lowering therapy.


All CHD patients participating in the study were surveyed by a cardiologist using medical history questionnaires. Before elective PCI, 64 CHD patients (N = 30 from SAP, N = 30 from UAP and N = 4 from AMI) had been taking a stable daily dose of atorvastatin 20 mg for 2–4 weeks; Before primary PCI, another 21 CHD patients had received simvastatin at 40 mg/day for three weeks (N = 5 from UAP and N = 16 from AMI). Blood samples were collected from subjects after a 12-h overnight fast and put into tubes containing 0.1 % EDTA.

The effect of HDL and total cholesterol (TC) levels on the HDL subclass distribution was assessed in all CHD patients based on a cut-off of 1.03 mmol/L for HDL and 5.17 mmol/L for TC. The reference levels of plasma HDL and TC were defined according to guidelines from the third Adult Treatment Panel guidelines of the National Cholesterol Education Program (NCEP-ATPIII) [18].

Angiographic Assessment of Coronary Arteries

Quantitative coronary angiography (QCA) was performed at the cardiology department of our hospital. QCA was carried out by resident physicians using the Judkins method, specifically through arteria radialis or arteria cruralis, multi-position projection. QCA was performed at baseline according to standard methods [19]. A minimum of 3 sets of orthogonal views of the left coronary artery and 1 of the right coronary artery were obtained from each subject. Analysis of angiograms was performed with a previously validated system of cine-projection. Briefly, the reference, minimal diameter (the point of greatest narrowing), and the average luminal diameters were obtained for 10 proximal epicardial coronary artery segments [20]. The mean minimal coronary artery diameter was calculated in each subject as the average of the minimal luminal diameter in the 10 coronary segments.

Anthropometric Data

Patients’ heights were measured in centimeters with shoes off and weights were measured in kilograms in indoor clothing. Body mass index (BMI) was calculated using the formula BMI = weight (kg)/height2 (m2).


Fasting (12 h) blood samples were collected in tubes containing 0.1 % EDTA prior to percutaneous intervention and centrifuged at 3,000 rpm for 20 min at 4 °C to obtain plasma. Plasma samples were stored at 4°C and used within 24 h for lipid and apolipoprotein analyses. Aliquots of each plasma sample were stored at −70 °C for the determination of HDL subclasses.

Plasma Lipid and Apolipoprotein Analyses

Plasma concentrations of TAG, TC, total HDL, and total LDL cholesterol, along with fasting plasma glucose, apoA-I, and apoB-100 values were measured for all of the subjects using automated standardized equipment by the Clinical Laboratory of West China Hospital, Sichuan University.

HDL Subclasses Analyses

The subclass distribution of HDL was determined with 2-dimensional (2-D) gel electrophoresis and subsequent immunodetection as described previously [21]. In brief, 10 μl of plasma was applied to 0.7 % agarose gel and electrophoresis in the first dimension. After electrophoretic separation of lipoproteins, further separation by electrophoresis was carried out along the 2–30 % nondenaturing polyacrylamide gradient gel in the second dimension. To determine HDL subclasses, Western blotting was conducted after 2-D gel electrophoretic, plasma proteins and molecular markers were electrophoretically transferred to polyvinylidene (PVDF) membranes, stained with 0.1 % Ponceau S, and the position of molecular standard protein bands was labeled with a pencil. They were then destained by diffusion and 5 % bovine serum albumin (BSA) was used to recover the membrane, followed by a reaction with horseradish peroxidase (HRP)-labeled goat antihuman apoA-I immunoglobulin G (IgG). The relative concentration of each HDL subclass was calculated as the percentage of total plasma apoA-I according to the density of each spot. HDL particle sizes were calibrated using a standard curve that included bovine serum albumin, ferritin, and thyroglobulin (Pharmacia Uppsala, Sweden). The relative percent concentration of each HDL subclass was multiplied by the apoA-I concentration in the sample to yield the relative concentrations of each HDL subclass of apoA-I (mg/L). The intra-assay variation (N = 5) of the specific HDL subclasses was 9.4 % (pre-β1-HDL), 9.8 % (pre-β2-HDL), 4.9 % (HDL3c), 6.2 % (HDL3b), 7.3 % (HDL3a), 11.1 % (HDL2a), and 7.9 % (HDL2b).

Statistical Analysis

The Kolmogorov–Smirnov test was used to determine the normality of distributions of plasma lipids and HDL subclasses. Non-Gaussian distribution was transformed to Gaussian distribution using the natural logarithm conversion. All values are presented as means ± standard deviation (SD). Significant differences between the groups were analyzed by one-way analysis of variance. Pearson correlation analysis was used to estimate the correlation between plasma lipoproteins and the changes in HDL subclass profile among CHD patients. A multiple regression analysis was also performed to examine the relationship between changes in coronary stenosis and the subclass composition of HDL. Differences were considered statistically significant at P < 0. 05. All statistical analyses were performed using the statistical package SPSS (Version 15.0, SPSS Inc).


Clinical Characteristics and HDL Subclass Distribution in the Different Types of CHD Subjects

Clinical characteristics for the SAP, UAP, and AMI subjects are shown in Table 1. In UAP and AMI subjects, BMI, waist circumference, waist to hip ratio, and the concentrations of HDL were lower compared to SAP subjects, but the levels of TC, LDL, fasting plasma glucose, and TAG, as well as the Gensini Score, were significantly higher than those in SAP patients. ApoA-I was significantly lower in AMI subjects than in SAP subjects. In addition, BMI, waist circumference, and waist to hip ratio were significantly lower, but TAG, TC, LDL, fasting plasma glucose levels, and the Gensini Score were higher in the subjects with AMI compared to those with UAP.
Table 1

The demographics, clinical characteristics and apoA-I contents of HDL subclasses among different types of CHD patients


SAP patients (N = 30)

UAP patients (N = 35)

AMI patients (N = 20)

CHD patients (N = 85)

Age (years)

64.2 ± 7.1

67.2 ± 7.8

66.8 ± 8.7

66.7 ± 7.7

BMI (kg/m2)

25.1 ± 4.9

24.6 ± 2.9

23.9 ± 2.4a†

24.8 ± 2.3

Waist (cm)

90.6 ± 6.2

90.9 ± 8.2

84.5 ± 5.8a*b*

90.4 ± 9.5

Hip (cm)

92.6 ± 6.1

92.8 ± 8.1

91.0 ± 5.3

92.6 ± 6.9


0.98 ± 0.1

0.99 ± 0.1

0.93 ± 0.1a†b†

0.98 ± 0.1

Gensini score

31.0 ± 6.3

50.8 ± 8.5a*

67.7 ± 11.9a†b*

47.9 ± 7.2

FPG (mmol/L)

6.1 ± 1.0

7.9 ± 1.2a†

9.9 ± 2.3a‡b†

7.8 ± 1.1

TAG (mmol/L)

1.9 ± 0.3

2.1 ± 0.2

2.9 ± 0.4a†b*

2.2 ± 0.4

TC (mmol/L)

4.0 ± 0.4

4.5 ± 0.8a†

4.9 ± 0.3a‡b*

4.4 ± 0.7

LDL (mmol/L)

3.1 ± 0.2

3.4 ± 0.3a*

4.0 ± 0.2a‡b†

3.3 ± 0.3

HDL (mmol/L)

1.2 ± 0.3

1.1 ± 0.2

1.0 ± 0.3a*

1.1 ± 0.2

ApoA-I (g/L)

1.3 ± 0.2

1.2 ± 0.3

1.1 ± 0.2a*

1.2 ± 0.3

ApoB-100 (g/L)

0.6 ± 0.1

0.7 ± 0.2

0.7 ± 0.1

0.7 ± 0.2

Preβ1-HDL (mg/L)

103.6 ± 7.8

110.9 ± 10.9

120.1 ± 11.3a*

110.7 ± 9.2

Preβ2-HDL (mg/L)

56.9 ± 6.3

55.7 ± 6.7

52.9 ± 5.6

55.5 ± 5.9

HDL3c (mg/L)

78.2 ± 9.7

89.8 ± 17.7

94.3 ± 20.2a*

90.7 ± 9.1

HDL3b (mg/L)

133.5 ± 17.6

141.6 ± 23.4

157.4 ± 36.3a*b*

140.8 ± 25.1

HDL3a (mg/L)

270.7 ± 46.8

289.7 ± 48.8a*

306.5 ± 51.3a†b*

285.1 ± 49.4

HDL2a (mg/L)

263.5 ± 40.9

228.8 ± 32.6a†

207.6 ± 28.7a‡b*

235.8 ± 34.5

HDL2b (mg/L)

344.0 ± 68.1

315.1 ± 58.2a†

225.5 ± 38.4a‡b‡

320.9 ± 57.7

Values are expressed as means ± SD

CHD coronary heart disease, HDL high density lipoprotein, SAP stable angina pectoris, UAP unstable angina pectoris, AMI acute myocardial infarction, BMI body mass index, TAG triacylglycerols, TC total cholesterol, LDL low density lipoprotein, HDL high density lipoprotein, ApoA-I apolipoproteinA-I, ApoB-100 apolipoproteinB-100, N number

aCompared with SAP patients

bCompared with UAP patients

P < 0.05,  P < 0.01,  P < 0.001

The contents of different HDL subclasses were then compared between the SAP, UAP, and AMI subjects (Table 1). The contents of HDL3a were significantly higher, while those of HDL2a and HDL2b were significantly lower, in UAP subjects compared to SAP subjects. Other HDL subclasses were not significantly different between these two groups. Furthermore, the AMI subjects had significantly higher contents of small-sized preβ1-HDL, HDL3c (SAP), HDL3b, and HDL3a, but reduced contents of large-sized HDL2a and HDL2b compared to SAP and UAP subjects.

Plasma Lipids, Apolipoprotein Levels and HDL Subclass Contents for CHD and Normolipidemic Control Subjects in Relation to Plasma LDL Levels

As shown in Table 2, the normolipidemic control subjects with LDL ≥3.34 mmol/L have higher levels of TAG, TC, LDL, ApoB-100, preβ1-HDL, HDL3a, and HDL3a, but lower HDL, ApoA-I, HDL2a, and HDL2b levels compared to normolipidemic control subjects with LDL <3.34 mmol/L. In consistent with this, in comparison with CHD with LDL <3.34 mmol/L, CHD with LDL ≥3.34 mmol/L subjects have higher levels of TAG, TC, LDL, ApoB-100, preβ1-HDL, HDL3 (HDL3c, HDL3b, and HDL3a); however, lower HDL, HDL2 (HDL2a, and HDL2b) contents.
Table 2

The plasma lipids, apolipoproteins levels, and HDL subclasses contents for CHD and normolipidemic control subjects in relation to plasma LDL levels


Normolipidemic control with LDL <3.34 (N = 68)

Normolipidemic control with LDL ≥3.34 (N = 12)

CHD with LDL <3.34 (N = 75)

CHD with LDL ≥3.34 (N = 10)

TAG (mmol/L)

1.0 ± 0.2

1.5 ± 0.2a†

2.0 ± 0.3b‡

2.6 ± 0.4a†b‡

TC (mmol/L)

4.8 ± 0.5

5.3 ± 0.4a†

4.1 ± 0.4b‡

4.7 ± 0.4a†b†

LDL (mmol/L)

2.8 ± 0.3

3.5 ± 0.4a‡

3.0 ± 0.3

3.7 ± 0.3a‡

HDL (mmol/L)

1.5 ± 0.2

1.0 ± 0.2a‡

1.2 ± 0.2b*

1.0 ± 0.2a*

ApoA-I (g/L)

1.6 ± 0.3

1.0 ± 0.2a‡

1.2 ± 0.2b†

1.1 ± 0.2

ApoB-100 (g/L)

0.5 ± 0.1

0.8 ± 0.2a†

0.6 ± 0.1

0.8 ± 0.2a*

Preβ1-HDL (mg/L)

72.1 ± 7.5

90.5 ± 8.7a*

100.8 ± 9.6b*

120.5 ± 10.3a*b†

Preβ2-HDL (mg/L)

50.2 ± 4.8

54.3 ± 5.0

52.5 ± 5.3

60.3 ± 5.6

HDL3c (mg/L)

63.4 ± 7.7

70.1 ± 8.4

83.9 ± 15.4b*

102.8 ± 23.1a*b†

HDL3b (mg/L)

116.5 ± 15.6

135.1 ± 22.1a*

133.5 ± 20.5b*

150.5 ± 25.3a*b*

HDL3a (mg/L)

236.3 ± 38.5

281.1 ± 41.2a†

275.1 ± 38.2b†

297.3 ± 37.6a*b*

HDL2a (mg/L)

317.7 ± 55.7

252.7 ± 40.1a‡

248.7 ± 33.1b‡

222.5 ± 34.4a*b†

HDL2b (mg/L)

412.9 ± 80.5

373.8 ± 70.4a†

347.6 ± 56.5b‡

302.3 ± 54.3a†b‡

Values are expressed as means ± SD

CHD coronary heart disease, TAG triacylglycerols, TC total cholesterol, LDL low density lipoprotein, HDL high density lipoprotein, ApoA-I apolipoproteinA-I, ApoB-100 apolipoproteinB-100, N number

aCompared with the corresponding low levels group

bCompared with the corresponding LDL levels among Normolipidemic Control Subjects

* P < 0.05, † P < 0.01,  P < 0.001

Additionally, both of CHD with LDL <3.34 mmol/L and CHD with LDL ≥3.34 mmol/L subjects had increased TAG, TC, preβ1-HDL, HDL3c, HDL3b, HDL3a, but decreased HDL (only in CHD with LDL <3.34 mmol/L), HDL2a, and HDL2b versus normolipidemic control subjects corresponding LDL levels, respectively.

Relations of the preβ1-HDL and HDL2b Contents to the Levels of LDL-C

To investigate the degree of HDL subclasses change with the levels of LDL-C varied, we using each LDL-C value as X axis, with every LDL-C corresponding the preβ1-HDL and HDL2b contents as Y axis to plot (Fig. 1). The figure shows associations of LDL-C levels with increased preβ1-HDL contents and with decreased HDL2b contents.
Fig. 1

Relations of the preβ1-HDL and HDL2b contents to the levels of LDL-C

The apoA-I Contents of HDL Subclasses for the Normolipidemic Control and CHD Subjects According to the Plasma TC Levels

As shown in Table 3, compared to normolipidemic control subjects with TC < 5.17 mmol/L, normolipidemic control subjects with a plasma TC ≥ 5.17 mmol/L had higher HDL3c, HDL3b, HDL3a, and preβ1-HDL, but significantly lower HDL2a and HDL2b contents.
Table 3

The ApoA-I contents of HDL subclasses among CHD patients and normolipidemic control subjects categorized by plasma TC levels


Normolipidemic control with TC <5.17 mmol/L (N = 65)

Normolipidemic control with TC ≥5.17 mmol/L (N = 15)

CHD with TC <5.17 mmol/L (N = 50)

CHD with TC ≥5.17 mmol/L (N = 35)

Preβ1-HDL (mg/L)

88.4 ± 8.5

103.5 ± 8.7a*

99.5 ± 9.2

118.4 ± 12.4a*b*

Preβ2-HDL (mg/L)

50.5 ± 4.6

51.8 ± 5.0

51.3 ± 5.8

53.4 ± 6.0

HDL3c (mg/L)

70.1 ± 7.4

90.1 ± 8.6a*

87.9 ± 10.2b*

104.9 ± 12.5a*b*

HDL3b (mg/L)

125.9 ± 15.6

140.1 ± 22.1a*

135.2 ± 25.3

147.8 ± 28.6

HDL3a (mg/L)

267.8 ± 40.5

308.9 ± 61.2a†

273.7 ± 26.3

295.9 ± 31.2a*

HDL2a (mg/L)

288.2 ± 35.7

257.2 ± 40.1a†

264.2 ± 31.2b*

215.0 ± 35.4a†b†

HDL2b (mg/L)

377.5 ± 70.8

334.5 ± 60.2a†

345.1 ± 68.1b†

303.0 ± 53.2a†b†

Values are expressed as means ± SD

CHD coronary heart disease, TC total cholesterol, HDL high density lipoprotein, ApoA-I apolipoproteinA-I, N number

aCompared with the corresponding low levels group

bCompared with the corresponding TC levels among Normolipidemic Control Subjects

* P < 0.05,  P < 0.01,  P < 0.001

Compared with CHD subjects with TC <5.17, CHD subjects with a plasma TC ≥5.17 mmol/L had lower HDL2a and HDL2b contents, but higher HDL3c, HDL3a and preβ1-HDL contents.

Moreover, CHD with TC >5.17 mmol/L and CHD with TC ≥5.17 mmol/L subjects had increased HDL3c and preβ1-HDL (only in CHD with TC ≥5.17 mmol/L); however, declined HDL2a and HDL2b compared with those in normolipidemic control subjects with corresponding levels, separately.

The apoA-I Contents of HDL Subclasses for the Normolipidemic Control and CHD Subjects in Accordance with the Plasma HDL Levels

Table 4 shows that in normolipidemic control subjects, the contents of preβ1-HDL, HDL3c, HDL3b, and HDL3a in the HDL ≥1.03 mmol/L subgroup were significantly lower than the HDL <1.03 mmol/L subgroup; in contrast, those of HDL2a and HDL2b in the HDL ≥1.03 mmol/L subgroup were significantly higher than in the HDL <1.03 mmol/L subgroup. Likewise, in CHD subjects, compared to the HDL <1.03 mmol/L group, the characteristics of the HDL subclasses distribution in HDL ≥1.03 mmol/L was in accord with that in normolipidemic control subjects with HDL ≥1.03 mmol/L.
Table 4

The apoA-I contents of HDL subclasses among CHD patients and normolipidemic control subjects categorized by plasma HDL levels


Normolipidemic control with HDL <1.03 mmol/L (N = 10)

Normolipidemic control with HDL ≥1.03 mmol/L (N = 70)

CHD with HDL <1.03 mmol/L (N = 17)

CHD with HDL ≥1.03 mmol/L (N = 68)

Preβ1-HDL (mg/L)

95.4 ± 10.4

75.1 ± 8.1a*

112.6 ± 10.6b*

91.9 ± 7.9a*b*

Preβ2-HDL (mg/L)

56.1 ± 4.8

50.6 ± 5.0

60.2 ± 5.1

51.2 ± 5.9

HDL3c (mg/L)

92.2 ± 8.4

70.0 ± 8.6a*

114.6 ± 14.9b*

84.8 ± 7.3a†b*

HDL3b (mg/L)

136.5 ± 19.6

115.4 ± 18.4a*

151.1 ± 25.4b*

130.8 ± 14.7a*b*

HDL3a (mg/L)

315.3 ± 40.5

247.2 ± 35.9a*

298.4 ± 38.7b*

265.6 ± 33.6a†b*

HDL2a (mg/L)

242.8 ± 40.1

278.6 ± 35.7a†

218.7 ± 33.0b‡

262.3 ± 33.2a†b*

HDL2b (mg/L)

327.6 ± 55.2

371.3 ± 60.8a†

303.0 ± 53.2b†

345.1 ± 68.1a†b*

Values are expressed as means ± SD

CHD coronary heart disease, HDL high density lipoprotein, ApoA-I apolipoproteinA-I, N number

aCompared with the corresponding low levels group

bCompared with the corresponding HDL levels among Normolipidemic Control Subjects

* P < 0.05, † P < 0.01,  P < 0.001

Furthermore, when compared with the normolipidemic control subjects with corresponding HDL levels, the CHD subjects in both HDL <1.03 mmol/L and HDL ≥1.03 mmol/L subgroups had higher contents of small preβ1-HDL, HDL3c, HDL3b, and HDL3a particles; however large sized HDL2 particles remained at the lower levels.

The Correlation Analysis Between Plasma Lipids, Lipoproteins, Fasting Plasma Glucose, and apoA-I Contents of HDL Subclasses

After measuring the fasting plasma glucose levels, correlation analysis results indicated that the levels of TC and HDL correlated positively with all HDL subclass contents. LDL levels showed a significant positive correlation with HDL3 but were inversely correlated with HDL2a and HDL2b. Moreover, TAG concentrations were positively correlated with preβ1-HDL, HDL3b, and HDL3a. When measuring the plasma TAG and TC levels, no significant correlation was found between plasma fasting plasma glucose levels and the HDL subclass contents (Table 5).
Table 5

Correlation coefficients between plasma lipids, FPG, and the ApoA-I contents of HDL subclasses in CHD patients

















































HDL high density lipoprotein, TAG triacylglycerols, TC total cholesterol, LDL low density lipoprotein, HDL high density lipoprotein cholesterol, CHD coronary heart disease, ApoA-I apolipoproteinA-I, FPG fasting plasma glucose

* P < 0.05, ** P < 0.01

Multivariate Analysis

We performed stepwise multivariate regression analyses using the Gensini Score as the dependent variable, and using plasma lipids, lipoproteins, and HDL subclasses as independent variables. The Gensini Score was significantly and independently predicted by HDL2b (standardized regression coefficient [SRC] −0.386, P < 0.05), apoA-I (SRC −0.271, P < 0.05), and HDL3b (SRC 0.245, P < 0.05). The adjusted r2 for this model was 0.390, implying that HDL2b, apoA-I, and HDL3b accounted for about 39 % of the variability in the Gensini Scores (Table 6).
Table 6

The Gensini score coefficients with plasma apolipoproteins levels, and HDL subclasses contents


Unstandardized coefficients

Standardized coefficients





Standard error


Gensini score



















HDL high density lipoprotein, ApoA-I apolipoproteinA-I, β beta


HDL are not a homogeneous category of lipoproteins, but consist of a set of subspecies with distinct structures, intravascular metabolism, and anti-atherogenic activities. By two-dimensional nondenaturing gel electrophoresis, immunoblotting, and image analysis methods, HDL can be subdivided into large-sized (HDL2a and HDL2b) and small-sized subclasses (preβ1-HDL, HDL3c, HDL3b, and HDL3a) and preβ2-HDL [20, 21].

Different HDL subclasses have distinct but complementary metabolic functions and the antiatherogenic actions of HDL reflect distinct functional properties of HDL particle subclasses rather than plasma levels of HDL [22]. Preβ1-HDL is referred to as an acceptor of cholesterol that effluxes from cells via the ATP binding cassette transporter A1 (ABCA1) in the first step of RCT [23]. Accumulation of small preβ1-HDL may be a result of inefficient conversion of preβ1-HDL to preβ2-HDL, or due to the esterification of cholesterol. It has been proposed that increases in preβ1-HDL concentrations reflect impairment in HDL maturation. Large cholesterol rich HDL2b particles may be important in determining the direction of the flow of cholesterol ester. When the HDL2b level is sufficient, most high-density lipoprotein-cholesterol ester is directed to the liver by the selective uptake of cholesterol ester by HDL receptors. However, in the absence of HDL2b particles, high-density lipoprotein-cholesterol ester is transferred to very low density lipoprotein (VLDL) and LDL by cholesterol ester transfer protein, resulting in an increased cholesterol ester in potentially atherogenic particles [24]. Asztalos et al. [25]. found that large α-1 HDL (HDL2b) was most significantly associated with CHD prevalence and each milligram per deciliter increase in a-1 HDL level decreased the odds of CHD by 26 %, using a model that included all established CHD risk factors. Therefore, the profile of HDL subclass distribution provides insight into CHD risk prediction, arteriographic progression, arteriographic outcome, and treatment response [26].

The treatment of patients with CHD has altered significantly over the past decade, driven by the prevalent use of HMG-CoA reductase inhibitors (collectively referred to as statins) [14], which inhibit the rate-limiting step in the cholesterol biosynthesis pathway. Treatment with HMG-CoA reductase inhibitors, such as simvastatin, during the study led to reductions in TC, LDL, and TAG averaging 20.8, 29.7, and 13.6 %, respectively, and a mean increase in HDL of 8 %. Most clinical studies have focused on increasing HDL levels, but several studies indicate that HDL levels can be dissociated from the cardioprotective functions of lipoprotein. A recent study was terminated prematurely because the rate of cardiovascular events increased when an agent that elevates HDL was added to statin therapy in established CHD subjects [2729]. Collectively, these observations indicate that alterations in HDL levels may not be the only determinant of the cardioprotective effects of HDL. Statins have also shown themselves to be effective in restoring the HDL subpopulation distribution, although with some variation, and this effect has been related to a change in total plasma lipids and a decrease in cholesterol ester transfer protein activity [30, 31]. In this setting, the reported variability in the individual response to statin treatment in terms of HDL subpopulations which may be related to individual genetic backgrounds and the cholesterol ester transfer protein TaqIB variant may also play a role [32].

Although increased CVD risk related to low HDL appears to be at least in part attributable to impaired antiatherogenic properties of this lipoprotein fraction, evidence is accumulating that HDL function, in terms of the modulation of inflammation and reverse cholesterol transport, is not accurately reflected by the plasma HDL cholesterol concentration as such [33]. On the other hand, the cholesterol content of HDL particles is mainly a historically founded and analytically feasible surrogate of RCT. HDL particles are not a homogenous class but can be further divided into subclasses. Most studies suggest that large HDL particles are associated with a favorable outcome, while small HDL particles may even be positively correlated with CV risk [34]. Correspondingly, the lipid disturbances of the metabolic syndrome include a decreased mean HDL particle size.

In the present study, we focused on the HDL subclass distribution phenotype in statin-treated CHD patients undergoing PCI. We found that although plasma lipids attained borderline-high (TAG) or optimal (TC and LDL) levels according to the NCEP-ATPIII guidelines [18], the distribution of HDL subclasses was abnormal in CHD patients undergoing PCI compared to those in normolipidemic healthy subjects. At the same time, to further dichotomize the CHD subjects using the cut-points 3.34 mmol/L for LDL, the levels of TAG, TC, and HDL (CHD with elevated LDL) as well as the distribution of HDL subclasses were abnormal in both of CHD with non-elevated LDL and CHD with elevated LDL patients who were undergoing PCI compared to those in normolipidemic healthy subjects. At the same time, compared to CHD patients with non-elevated LDL, CHD patients with elevated LDL had increased small-sized particles, but decreased large-sized particles. Further, for grouped analyses, individual baseline LDL value as X axis, with every LDL corresponding the preβ1-HDL and HDL2b contents as Y axis. Trends in mean values of major HDL subclasses (preβ1-HDL and HDL2b) across the each LDL were assessed through simple linear regression, in this model with the contents of preβ1-HDL and HDL2b as the dependent variable and the levels of LDL levels as the independent variable, our results revealed that associations of LDL levels with increased preβ1-HDL contents and with decreased HDL2b contents; however, these is not a better liner association between HDL subclasses and LDL levels. When compared to the normolipidemic control group, CHD patients with lower LDL levels have HDL3b levels similar to the control group. Although the HDL2b levels were lower than controls, they were significantly higher than the CHD group with LDL ≥3.34 mmol/L. On the basis of the above data, it is indicated that differences in these HDL subclasses did not simply reflect differences in LDL levels.

Our work also demonstrates that more than half of the CHD patients who were undergoing PCI had non-elevated TAG and TC levels. For these patients, HDL subclass distributions have been modified, including reduced levels of large-sized HDL2 particles and increased levels of small-sized preβ1-HDL particles. Similarly, HDL subclasses had not been normalized in CHD subjects with normal HDL levels.

Patients show individual differences in their response to the lipid-lowering treatment, and furthermore, many patients still die or have complications from cardiovascular events in spite of the lowering of plasma cholesterol levels. Despite the fact that statins effectively reduce plasma TC, LDL, and TAG levels, recent data suggest that statin treatment is not sufficient as a monotherapy to normalize the HDL subclasses distribution phenotype in patients with CHD. Consequently, potential residual cardiovascular risk remains high among the patients with abnormal HDL subclass distribution. Given these conditions, we predict that a test of HDL subclasses distribution might be a better index of cardiovascular health risk and useful for the further guiding of treatment strategies. Green et al. [35]. found that combination therapy with statin and niacin altered the protein composition of HDL3 to resemble more closely that in control subjects. Their observations raise the possibility that monitoring the protein composition of HDL could provide a measure of the therapeutic efficacy of lipid interventions.

The Pearson correlation results support the hypothesis that statins modify the components of the lipid profile, often to a greater degree than their impact on HDL subclass distribution. In addition, the multiple stepwise regression analysis reveals that the Gensini Score can be independently predicted by HDL2b and HDL3b. Although the levels of fasting plasma glucose were significantly higher in patients with CHD, after controlling for plasma TAG and TC levels, no significant correlation with plasma fasting plasma glucose levels and HDL subclasses was found.

Furthermore, when we divided these CHD patients into three subgroups according to condition type (SAP, UAP, and AMI groups), we found that the generation of small-sized HDL particles increased more markedly in the AMI subjects. The Gensini Score was significantly, independently, and inversely correlated with HDL2b, and independently and positively correlated with HDL3b. This finding is in agreement with Johansson and colleagues [36].

Our findings indicate that treatment with statins is effective for modifying plasma lipid levels, but not adequate as a monotherapy to normalize the HDL subclass distribution profile of CHD patients. Results from these studies demonstrated that the HDL subclass distribution may aid in risk stratification, especially in patients with CHD and therapeutic LDL and HDL levels.

In the current work, we mainly discussed the HDL subclass distribution profile among statin-treated CHD patients undergoing PCI. Given that many patients included in the study are likely to be on a lipid-lowering therapy (which is probably not standardized) with short treatment periods. In addition, baseline measurements of lipid profiles were not obtained prior to the initiation of statin therapy and clinical outcomes were not measured, we intend to perfect the study further.


This work was supported by the grants from the Fundamental Research Funds for the Central Universities (Grant No. 2010SCU11029) and China Postdoctoral Science Foundation (Grant No. 20110491719).

Conflict of interest

The authors have declared no conflicts of interests.

Copyright information

© AOCS 2012