HDL Levels and Particle Size: Does Size Matter at Both Ends of the Scale?

Linda Brookes, MSc

August 07, 2007

Raising HDL Levels

It is generally accepted that plasma levels of high-density lipoprotein (HDL) cholesterol are inversely related to the incidence of coronary heart disease (CHD) and stroke, and that therapeutically raising HDL cholesterol levels may therefore serve as a treatment for atherosclerotic vascular disease. The possibility of therapeutically raising HDL levels is especially important as it has become apparent that the current best therapy for atherosclerotic disease, lowering low-density lipoprotein (LDL) cholesterol levels with statins, still leaves a large percentage of residual risk. (That is, in clinical trials a significant percentage of patients on the best possible statin therapy still experience events.)

Although several approaches to raising HDL levels have been under active research, the most promising new class of drugs was based on inhibition of the cholesteryl ester transfer protein (CETP), with several pharmaceutical companies actively researching such agents. The concept received a major setback at the end of 2006, however, when development of the first drug in the class, torcetrapib, was terminated following its early failure in a large clinical trial, the Investigation of Lipid Level management to Understand its iMpact IN ATherosclerotic Events (ILLUMINATE).[1]

To review briefly, ILLUMINATE[2] was a large trial that enrolled 15,000 patients at high risk for CHD; it was terminated early by the Data Safety and Monitoring Board when interim analysis showed an excess of deaths, myocardial infarction (MI), angina, revascularization procedures, and heart failure in patients receiving torcetrapib plus atorvastatin compared with those receiving atorvastatin alone. These adverse clinical outcomes occurred despite unprecedented improvement in the relevant serum cholesterol values, with mean decreases and increases in LDL and HDL cholesterol, respectively, of approximately 10% to 15% and 55% to 60%.

The ILLUMINATE data are currently being analyzed, but in the meantime, possible reasons for the failure of the CETP inhibitor are being actively researched. One of the principal questions, of course, has been a re-evaluation of whether CETP is a valid target for raising HDL cholesterol levels; as a further possibility it has even been suggested that high levels of HDL cholesterol might actually have an adverse effect on CHD risk. As counterintuitive as this may seem, 2 important presentations at the 2007 European Atherosclerosis Society (EAS) conference presented the results of investigations into precisely this possibility. These 2 studies suggested that very high levels of HDL cholesterol may not be associated with reduced cardiovascular risk and that very large HDL particle size may not confer protection against cardiovascular disease.

IDEAL Clues?

Dr. Holme and his colleagues, at the Ullevål University Hospital, Oslo, Norway, reported an investigation of the relationship between high HDL cholesterol levels and CHD event rates that suggested that, with adjustment for apolipoprotein (apo)A-1 and other lipoprotein variables, very high levels of HDL cholesterol may be a poor marker of cardioprotection, whereas apoA-1 may be associated with reduced risk for major cardiovascular events, with or without adjustments for HDL cholesterol and apoB.[3] These results came from a post hoc analysis of data from the Incremental Decrease in Events through Aggressive Lipid Lowering (IDEAL) trial.

The IDEAL trial, which was also sponsored by Pfizer (New York, NY; the manufacturer of torcetrapib and supporter of the ILLUMINATE trial), was a prospective, randomized, open-label, blinded endpoint evaluation (PROBE) trial conducted at 190 ambulatory cardiology care and specialist practices in Denmark, Finland, Iceland, The Netherlands, Norway, and Sweden between March 1999 and March 2005.[4] The trial enrolled 8888 patients aged ≤ 80 years with a history of MI and randomized them, without washout from existing statin therapy, to receive either intensive statin (atorvastatin 80 mg) or standard statin (simvastatin 20-40 mg) treatment. Over a median follow-up of 4.8 years, the mean LDL cholesterol level in the atorvastatin group was 23 mg/dL (0.6 mmol/L) lower than in the simvastatin group.

Intensive treatment with atorvastatin in IDEAL reduced the risk for major coronary events (CHD death, nonfatal MI, and resuscitated cardiac arrest), the primary endpoint of the trial, by 11% compared with simvastatin (hazard ratio [HR] = 0.89, 95% confidence interval [CI] 0.78, 1.01; P = .07). Intensive treatment with atorvastatin was also associated with a 13% reduction in the risk for major cardiovascular events (major coronary events plus stroke) (HR = 0.87, 95% CI 0.78, 0.99; P = .02), a 16% reduction in the risk for any CHD event (HR = 0.84, 95% CI 0.76, 0.91; P < .001), and a 16% reduction in the risk for any cardiovascular event (HR = 0.84, 95% CI 0.87, 0.91; P < .001) compared with simvastatin.

In a separate presentation at the EAS meeting,[5] Dr. Holme and colleagues presented the results of an analysis of on-treatment HDL cholesterol levels and the risk for major cardiovascular events in the IDEAL trial. The method here was to calculate major cardiovascular events according to quintiles of on-treatment HDL cholesterol levels (averages at months 3 and 6). (However, adjustments for apoA-1 and apoB were not made in this analysis.)

As expected, on-treatment HDL cholesterol levels remained predictive of cardiovascular disease risk, with an inverse relationship seen between the incidence of major cardiovascular events and on-treatment HDL cholesterol levels (HR 0.991, 95% CI 0.86, 0.96; P = .0009). On-treatment HDL cholesterol levels were predictive of major cardiovascular event incidence across the IDEAL cohort, with a 1-mg/dL (0.026- mol/L) increment in HDL cholesterol concentration corresponding with an almost 1% (0.83%) reduction in relative major cardiovascular event risk. Lower on-treatment HDL levels remained predictive of major cardiovascular event frequency when the trial cohort was divided into 2 LDL cholesterol subgroups (> 80 mg/dL and ≤ 80 mg/dL), suggesting that HDL cholesterol levels may be an important consideration, even in patients whose LDL cholesterol levels are ≤ 80 mg/dL.

For their analysis, Dr. Holme and his colleagues were particularly interested in the CHD event rates in those IDEAL patients who had the highest levels of HDL cholesterol, especially > 70 mg/dL (> 1.8 mmol/L), in the light of the findings with torcetrapib treatment in ILLUMINATE, which had resulted in similar high levels of HDL cholesterol. However, in contradistinction to the previous study, they also evaluated the predictive ability of increased apoA-1, the primary protein constituent of HDL, on the risk for major cardiovascular events when HDL cholesterol and apoB, the primary apolipoprotein of LDL, were kept constant.

Data from the 8888 IDEAL patients -- excluding those who had a major cardiovascular event during the first 6 months of the trial, but including 679 patients who had a major cardiovascular event after the first 6 months -- were evaluated with Cox fixed covariate regression analysis with time to a major cardiovascular event after 6 months as a dependent variable. All models were adjusted for age, sex, and smoking.

After adjusting for apoB and apoA-1 levels (ie, keeping them constant while looking only at variations in HDL levels), they found that on-treatment HDL cholesterol levels were indeed a significant risk factor for a major cardiovascular event (P = .038) ( ), indicating that high HDL cholesterol levels could actually be an adverse risk factor. Adjustment for LDL cholesterol did not alter the relationship significantly.

  HDL Cholesterol vs Major Cardiovascular Events

Adjustment Variable HR 1 SD Increase 95% CI P Value
None 0.92 0.85, 1.00 .043
LDL cholesterol 0.91 0.84, 0.99 .021
apoA-1 1.05 0.96, 1.15 .59
apoA-1, LDL cholesterol 1.11 0.93, 1.33 .26
apoA-1, apoB 1.21 1.01, 1.46 .038
apoA-1, LDL cholesterol, apoB 1.26 1.04, 1.52 .020

HDL = high-density lipoprotein; LDL = low-density lipoprotein; HR = hazard ratio; CI = confidence interval
All models adjusted for age, sex, and smoking. 1 SD increase in HDL cholesterol = 12 mg/dL (0.31 mmol/L)

When both treatment arms in the IDEAL trial were combined, Dr. Holme noted that only 30 major cardiovascular events occurred in patients with HDL cholesterol > 70 mg/dL (1.8 mmol/L) ( ).

  HDL Cholesterol Levels (Treatment Groups Combined)

HDL Cholesterol (mg/dL)* Patients (n) MCE (n)
< 40 2583 222
40-50 2495 238
50-60 2059 145
60-70 650 44
70-80 215 20
> 80 112 10

*On-treatment levels (averages of months 3 and 6)
HDL = high-density cholesterol; MCE= major cardiovascular event

After adjustment for apoB and apoA-1, however, the increased risk was particularly evident in patients with on-treatment HDL cholesterol levels > 70 mg/dL; these patients had a relative risk of 2.23 (P = .014) compared with patients with HDL cholesterol < 40 mg/dL. In contrast, higher apoA-1 levels were associated with decreased risk, whether or not adjustments were made for HDL cholesterol and apoB ( ).

  Relationship of apoA-1 to Risk for Major Cardiovascular Events (Adjusted for apoB and HDL Cholesterol)

Adjustment Variable HR 1 SD Increment 95% CI P Value
None 0.90 0.83, 0.98 .012
HDL cholesterol 0.86 0.71, 1.03 .108
HDL cholesterol, apoB 0.74 0.61, 0.90 .002

All models adjusted for age, sex, and smoking; 1 SD increase in apoA-1 = 0.22 g/L
HDL = high-density lipoprotein; HR = hazard ratio; CI = confidence interval

Dr. Holme acknowledged the limitations of a post hoc analysis in which adjustments were not made for lifestyle factors, concomitant medications, or other medical considerations. "These are observational findings from the treatment groups combined and not from randomized comparisons," he stressed. Nevertheless, the result that the highest HDL cholesterol levels (> 70 mg/dL), but not high apoA-1 levels, correlate with increased cardiovascular risk clearly warrants further attention.

Relationship Between HDL Particle Size and Coronary Artery Disease Risk

One of the effects of CETP inhibition is an increase in HDL particle size, and because the adverse effects of torcetrapib became apparent, it has been hypothesized that the increase in size induced by CETP inhibition might affect the antiatherogenic capacity of the HDL particles, resulting in less functional or even dysfunctional HDL. In another analysis presented at the 2007 EAS meeting, it was suggested that, after adjustment for apoA-I and apoB levels, very large HDL particles would be associated with an increased risk for coronary artery disease. This observation may have consequences for the development of novel compounds targeting HDL, Dr. van der Steeg predicted.[6]

He and his colleagues carried out a case-control study of participants in the European Prospective Investigation into Cancer and Nutrition (EPIC)-Norfolk. EPIC-Norfolk was a prospective population study of 25,663 men and women aged 45-79 years residing in Norfolk, United Kingdom, who completed a baseline questionnaire survey and attended a clinic visit. Participants were identified from age and sex registers of general practices in Norfolk as part of the 10-country collaborative EPIC survey, which was designed to investigate dietary and other determinants of cancer.[7] Additional data were obtained in EPIC-Norfolk so that determinants of other diseases could also be assessed.

At the baseline survey between 1993 and 1997, participants completed a detailed health and lifestyle questionnaire. Nonfasting blood samples were obtained, and stored blood samples were processed for assay at the Department of Clinical Biochemistry, University of Cambridge, Cambridge, United Kingdom, or were stored at -80°C. All participants were flagged for death certification at the United Kingdom Office of National Statistics, and vital status was ascertained for the entire cohort. Hospitalized participants were identified by using their unique National Health Service.

For their analysis, Dr. van der Steeg's team used data from a case-control study carried out among 858 cases of fatal or nonfatal MI compared with 1491 matched controls with no statin use and no history of MI at the time of enrollment. They then used nuclear magnetic resonance (NMR) spectroscopy to assess the relationship between HDL particle size in the stored blood samples and the occurrence of cardiovascular disease, focusing on the effect of very large HDL particle size.

To assess the relationship with HDL particle size, conditional logistic regression analysis was used, taking into account the matching controls (basic model), matched for age, sex, and enrollment period, and adjusted for classic cardiovascular risk factors, including smoking (yes/no), body mass index, diabetes mellitus, systolic blood pressure, and alcohol intake (units per week). Additional adjustment was made for proatherogenic lipoproteins, including apoB and apoA-1.

The odds ratio (OR) per 1 SD increase in HDL particle size for each subgroup was calculated, with the OR for the first subgroup defined as 1.00. As seen in , the number of major cardiovascular events decreased with increasing HDL particle size.

  Subgroup Definition

Subgroup HDL Size (mm) Patients (%) MCE (n)
1 < 8.5 22.5 229
2 8.5-9.0 38.3 356
3 9.0-9.5 25.8 182
4 9.5-10.0 10.9 79
5 > 10.0 2.0 12

MCE = major cardiovascular event

As expected, initial regression analysis showed that each 1 SD increase in HDL particle size was associated with a significantly deceased risk. However, after adjustment for apoA-I and apoB levels, this relationship changed to a significantly increased risk ( ). The significantly increased risk was limited to the highest subgroups (OR 1.77 [95% CI 1.13-2.77] and 2.57 [1.13-5.85], for subgroups 4 and 5 compared with subgroup 1, respectively).

  HDL Particle Size and MI Risk

Adjustment Variable OR per 1 SD 95% CI P Value
Classic cardiovascular risk factor* 0.87 0.78, 0.97 .009
+ adjustment for apoA-1 and apoB 1.23 1.06, 1.42 .005

*Age and sex (by matching), smoking status (current/former/never), body mass index, diabetes mellitus, systolic blood pressure, and alcohol consumption (units/week); 1 SD increase of HDL size, +0.48 nm
HDL = high-density lipoprotein; MI = myocardial infarction; OR = odds ratio

A similar analysis by HDL cholesterol levels (as opposed to particle size) showed a reduced risk for MI with each 1 SD increase in HDL cholesterol level, but this became nonsignificant after adjustment for apoA-1 and apoB. Similar to the IDEAL analysis above, a slight trend was seen toward increased risk being associated with the highest HDL cholesterol levels (> 88.9 mg/dL) ( ).

  HDL Cholesterol and MI Risk

Adjustment Variable OR per 1 SD P Value
Classic cardiovascular risk factor* 0.79 < .0001
+ adjustment for apoA-1 and apoB 1.07 .46

*Age and sex (by matching), smoking status (current/former/never), body mass index, diabetes mellitus, systolic blood pressure, and alcohol consumption (units/week); 1 SD increase of HDL cholesterol, +15.2 mg/dL
HDL = high-density lipoprotein; MI = myocardial infarction; OR = odds ratio

Analysis by apoA-1 levels showed a different relationship, with a significantly reduced risk and no difference at the higher levels of apoA-1 (≤ 235 mg/dL) ( ).

  ApoA-1 and MI Risk

Adjustment Variable OR per 1 SD P Value
Classic cardiovascular risk factor* 0.79 < .0001
+ adjustment for apoA-1 and apoB 0.74 .001

*Age and sex (by matching), smoking status (current/former/never), body mass index, diabetes mellitus, systolic blood pressure, and alcohol consumption (units/week); 1 SD increase of apoA-1, +30 mg/dLMI = myocardial infarction; OR = odds ratio

These data suggest that very large HDL particles no longer seem to confer protection against cardiovascular disease, at least when levels of apoA-1 and apoB are kept constant, but this lack of protection does not seem to hold true for apoA-1, Dr. van der Steeg concluded. Dr. van der Steeg revealed that he and his colleagues are currently analyzing data on the concentrations of large and small HDL particles in this group, which may add further information about the effects of HDL particle size.

The EPIC-Norfolk study is supported by the Medical Research Council (United Kingdom) and Cancer Research UK, and receives additional support from the European Union, Stroke Association, British Heart Foundation, United Kingdom Department of Health, Food Standards Agency, and the Wellcome Trust.


  1. In interests of patient safety, Pfizer stops all torcetrapib clinical trials; company has notified FDA and is in the process of notifying all clinical investigators and other regulatory authorities [press release]. New York: Pfizer; December 244, 2006.

  2. Nissen SE, Tardif JC, Nicholls SJ, et al; ILLUSTRATE Investigators. Effect of torcetrapib on the progression of coronary atherosclerosis. N Engl J Med. 2007;356:1304-1316.

  3. Holme I, Kastelein JJP, Faergeman O, et al. High HDL-C and CHD risk in statin-treated CHD patients: analysis of Incremental Decrease in Endpoints through decrease in endpoints through Aggressive Lipid-lowering (IDEAL) trial. Atherosclerosis. 2007;8:227. Abstract LB-OR-2.

  4. Pedersen TR, Faergeman O, Kastelein JJ, et al; Incremental Decrease in End Points Through Aggressive Lipid Lowering (IDEAL) Study Group. High-dose atorvastatin vs usual-dose simvastatin for secondary prevention after myocardial infarction: the IDEAL study: a randomized controlled trial. JAMA. 2005;294:2437-2445.

  5. Holme I, Gaffney M, Cater NB; IDEAL Trial Steering Committee and Investigators. On-treatment HDL-C levels and risk of major cardiovascular events in the IDEAL trial. Atherosclerosis. 2007;8:197. Abstract PO23-741.

  6. van der Steeg WA, Holme I, Boekholdt SM, et al. Relationship between high-density lipoprotein particle size and coronary artery disease risk in the EPIC-Norfolk prospective study. Atherosclerosis. 2007;8:227. Abstract LB-OR-3.

  7. Day N, Oakes S, Luben R, et al. EPIC-Norfolk: study design and characteristics of the cohort. European Prospective Investigation of Cancer. Br J Cancer. 1999;80(suppl1):95-103.