Small, Dense Low-Density Lipoprotein Particles as a Predictor of the Risk of Ischemic Heart Disease in Men

Although the association between LDL cholesterol levels and IHD is well accepted, a relatively high proportion of patients with IHD have plasma LDL cholesterol levels in the normal range.¹ LDL particles are heterogeneous with respect to size, density, composition, and physicochemical properties.²³⁴⁵ Several approaches have been used to characterize LDL particles. Using gradient gel electrophoretic analysis of isolated LDL, Austin et al⁶ dichotomized LDL particles into two distinct LDL phenotypes: pattern B, with a predominance of small, dense LDL particles, and pattern A, with a higher proportion of large, more buoyant LDL particles. Others have developed a global score representative of the distribution of LDL particle size and density values.⁷⁸⁹¹⁰ Irrespective of the approach used to characterize LDL particles and of the case definition, dense LDL particles were more prevalent among IHD case patients than among IHD-free control subjects.^10111213 In some instances, the relationship between the dense LDL phenotype and IHD remained significant after control for age, degree of obesity, and sex.¹⁴¹⁵

An increased proportion of small, dense LDL particles has also been associated with marked alterations in plasma lipoprotein and lipid levels, such as elevated TG and apo B concentrations and reduced HDL cholesterol levels, all of which are highly predictive of an increased risk for IHD.^81013141516 Whether the increased IHD risk associated with the presence of small, dense LDL is independent of the concomitant variation in plasma lipid levels remains to be clearly established, particularly through prospective studies.¹⁷ The respective contributions of the dense LDL phenotype and of lipoprotein-lipid levels to the subsequent development of IHD were examined in a subsample of men involved in the prospective phase of the Que´bec Cardiovascular Study. The LDL phenotype was determined by measurement of LDL-PPD in men initially free of clinical manifestations of IHD at entry, who eventually developed IHD over a 5-year follow-up. These men were compared with a group of matched control subjects who remained free of IHD during the same period.

Methods

Study Population and Follow-up

The Que´bec Cardiovascular Study cohort has been described in detail previously.¹⁸¹⁹ Briefly, in 1973 a random sample of 4637 men 35 to 64 years old was recruited by use of the provincial electoral lists from seven counties of the Que´bec metropolitan area for an evaluation of cardiovascular risk factors. Subsequent evaluations were performed at regular intervals, and data collected in 1985 were used as the baseline characteristics for the present prospective analyses. At the 1985 evaluation, 61% (n=2443) of the living cohort came to the lipid clinic in a fasting state for their evaluation. Among the 1557 other living potential participants, 10% could not be located, 19% showed up in a nonfasting state, and 71% either refused to participate or were evaluated in a nonfasting state at their homes by nurses of the project. Analyses of data collected in 1973 revealed that the age distribution of the 2443 participating men in 1985 was representative of the original cohort. In 1990 to 1991, all participants were contacted by mail and invited to answer a short standardized questionnaire on smoking habits, medication use, and history of cardiovascular diseases and diabetes mellitus. For those who reported such diseases and those who had died, hospital charts were reviewed. Telephone calls were made to participants who did not answer a second letter or, if unsuccessful, to a close family member. Mortality and morbidity data were obtained in 99% and 96%, respectively, of the participants of the 1973 initial screening. No difference in the frequency of IHD mortality was observed between the participants and the nonparticipants.

Evaluation of Risk Factors

Data on demographic and lifestyle variables as well as medical history and medication were obtained in 1985 through a standardized questionnaire administered to each participant by trained nurses and further reviewed by a physician. Body weight and height were recorded. Resting blood pressure was measured after a 5-minute rest in a sitting position. The mean of two blood pressure measures taken 5 minutes apart was used in the analyses. The following data were compiled from the questionnaire: personal and family history of IHD and diabetes, smoking habits, alcohol consumption, and medication use. Diabetes was considered in men who self-reported the disease or who were treated with hypoglycemic agents. Only 2% of men were using hypolipidemic drugs in 1985, whereas 8% and 4% of men, respectively, were using β-blockers and diuretics on a regular basis at the 1985 screening. Alcohol intake was computed from the type of beverage (beer, wine, and spirits) consumed in ounces per week and then standardized as absolute quantity, 1 oz of absolute alcohol being equivalent to 22.5 g of alcohol. Family history of IHD was considered positive if at least one parent and/or one sibling had a previous history of IHD.

Definition of IHD Events

The diagnosis of a first IHD event included typical effort angina, coronary insufficiency, nonfatal myocardial infarction, and coronary death. All myocardial infarction case patients met the criteria previously described,¹⁸ namely, diagnostic ECG changes alone or two of the following criteria: typical chest pain of ≥20 minutes' duration, abnormal creatine kinase enzyme at least twice the upper limit of normal, or characteristic ECG changes. Coronary insufficiency was considered if typical retrosternal chest pain of at least 15 minutes' duration was associated with transient ischemic ECG changes but without significant elevation in levels of creatine kinase. Diagnoses of all myocardial infarctions and coronary insufficiency were confirmed by hospital charts. All ECGs were read by the same cardiologist, who was unaware of the participants' risk profile. The diagnosis of effort angina was based on typical symptoms of retrosternal squeezing or pressure-type discomfort occurring on exertion and relieved by rest and/or nitroglycerin. Criteria for the diagnosis of coronary deaths included confirmation from death certificates, or autopsy report confirming the presence of coronary disease and without evidence for noncardiac disease that could explain death. Myocardial infarction was considered fatal if death occurred within 4 weeks of the initial event or if it was diagnosed at autopsy. IHD-related deaths were confirmed from the Provincial Death Registry. Informed consent was obtained to review relevant hospital files. Autopsies were performed in about one third of deaths.

Pairing Procedures

From the sample of 2103 men without clinical evidence of IHD in 1985 and with a complete profile, 114 developed IHD during the 5-year follow-up that ended in September 1990.¹⁹ The 114 cases of IHD included 48 first cases of myocardial infarction, 51 cases of effort angina, and 15 IHD-related deaths. Each subject with confirmed IHD (case subject) was matched with a control subject selected among the remaining 1989 men who showed no clinical signs of IHD during follow-up.²⁰ Subjects were matched on the basis of age, cigarette smoking, body mass index, and weekly alcohol intake. The mean within-pair differences for matching were 0.6 year, 0.2 kg/m², and 0.2 oz/wk for age, body mass index, and alcohol intake, respectively. The mean within-pair difference for cigarette smoking was 0.3 cigarette per day. A total of 11 pairs of men were excluded because of the impossibility of measuring LDL size or because they could not be matched owing to extreme values of cigarettes smoked per day.

Laboratory Analyses

Twelve-hour fasting blood samples were obtained from an antecubital vein while participants were in a sitting posture. A tourniquet was used but was released before blood withdrawal into vacuum tubes (Vacutainer, Becton-Dickinson) containing EDTA. Plasma was separated from blood cells by centrifugation and was immediately used for lipid and apolipoprotein measurements. Aliquots of fasting plasma were frozen at the time of collection and were later used for the assessment of LDL size. Plasma cholesterol and TG levels were determined on an Auto Analyzer II (Technicon Instruments Corp) as previously described.²¹ HDL cholesterol was measured in the supernatant fraction after precipitation of apo B–containing lipoproteins with heparin–manganese chloride.²² LDL cholesterol levels were estimated by the equation of Friedewald et al,²³ since men with TG levels >4.5 mmol/L were excluded from the analyses. Plasma apo B levels were measured by the rocket immunoelectrophoresis method of Laurell²⁴ as previously described.²¹ The coefficients of variation for cholesterol, HDL cholesterol, TG, and apo B measurements were all <3%.

Determination of LDL-PPD

Nondenaturing 2% to 16% PAGE was performed on whole plasma according to the procedures described by Krauss and Burke² and McNamara et al.¹⁶ Gels were prepared in our laboratory as previously described.²⁵ Aliquots of 7.5 μL of plasma samples were applied on gels in a final concentration of 20% sucrose and 0.25% bromphenol blue. After a 15-minute prerun, electrophoresis was performed at 200 V for 12 to 16 hours and at 400 V for 2 to 4 hours. Gels were stained according to standardized procedures and stored in a 9% acetic acid/20% methanol solution until analysis with an optical densitometric image analyzer (BioImage Visage 110) coupled to a SPARC Station 2 Sun computer (Millipore) and GEL 1D software. LDL-PPD was obtained from the migration of standards of known diameter such as ferritin, thyroglobulin, and 38.0-nm latex beads (Duke Scientific Corp) and plasma standards kindly provided by Dr R.M. Krauss. The estimated diameter for the major peak in each scan was identified as the LDL-PPD. One assay of LDL-PPD was performed for each subject. Analyses of pooled plasma standards revealed that the identification of the major LDL peak was highly reproducible, with an interassay coefficient of variation of <3% (unpublished data).

Statistical Analyses

LDL-PPD was measured in 103 case-control pairs. Baseline characteristics of men who developed IHD during the 5-year follow-up were compared with the characteristics of those who remained IHD-free by Student's t test for means and χ² for frequency data. Variables with a skewed distribution were log-transformed. Associations among variables were assessed with the Pearson and Spearman correlation coefficients for parametric and nonparametric variables, respectively. Conditional logistic regression analysis was used to assess the association between risk factors and IHD. Preliminary analyses have shown that the relationship of plasma lipid, lipoprotein, and apo B to IHD was essentially linear across their distribution. These variables were therefore treated as continuous in logistic models. The association between LDL-PPD and IHD was investigated with the variable as continuous and also by categorization of the LDL-PPD distribution into three groups with the tertile values of the control group (25.64 and 26.05 nm) as cutoff points. The risk of IHD associated with each continuous variable was standardized as the relative odds of IHD for a 1-SD increase in the concentration of the variable. Risk of developing IHD during the 5-year follow-up period in men having small (LDL-PPD≤25.64 nm) and intermediate (25.64<LDL-PPD≤26.05 nm) LDL-PPDs at baseline was estimated as the relative risk of IHD compared with men having large LDL particles (LDL-PPD>26.05 nm) at baseline. Odds were adjusted for the potential confounding effects of diabetes mellitus, medication use, family history of IHD, and systolic blood pressure. Diabetes mellitus, medication use, and family history were treated as categorical variables (presence or absence), whereas systolic blood pressure was treated as continuous. Statistical analyses were all performed on SAS software (SAS Institute).

Results

Baseline characteristics of men who developed IHD (case patients) and who remained IHD-free during the five-year follow-up (control subjects) are presented in Table 1. Mean age was 59±7 years, and the frequency of smokers in case patients and control subjects was identical (42%). The study design also eliminated the case-control differences in body mass index and alcohol consumption. Diabetes mellitus was more prevalent in 1985 in case patients than in control subjects (14.6% versus 1%, P<.001). The baseline prevalence of men using β-blockers and/or diuretics on a regular basis was also higher in case patients than in control subjects (18% versus 8%, P=.02). Plasma concentrations of cholesterol, LDL cholesterol, TGs, and apo B were significantly increased in case patients compared with control subjects. Case patients were also characterized by a trend for reduced HDL cholesterol levels compared with control subjects and by a higher mean total/HDL cholesterol ratio (P=.002). Systolic blood pressure was significantly higher in case patients than in control subjects (137 versus 132 mm Hg, P=.05). Although there was a trend for smaller mean LDL-PPD in case patients than in control subjects, the difference did not reach statistical significance (P=.16). The lack of difference between case patients and control subjects in mean LDL-PPD could be attributed to the fact that a few control subjects had very low LDL-PPD (<24.2 nm). As shown in Fig 1, however, the distribution of LDL-PPD in men who did not develop IHD during follow-up tended to be shifted toward higher values compared with the IHD case patients (χ²P=.03). By definition, one third of control subjects were found in each tertile of LDL-PPD. A smaller proportion of case patients (19.4%) were found within the third tertile of the LDL-PPD distribution (large LDL particles). Similar proportions of IHD case patients and control subjects were found in tertile 2, whereas the proportion of men having LDL-PPD≤25.64 nm was higher in case patients than in control subjects (49.5% versus 34%). As shown in Fig 2, LDL-PPD showed relatively strong associations with TG (r=−.46, P<.001), HDL cholesterol (r=.39, P<.001), and apo B levels (r=−.25, P<.001) but showed no relation to total cholesterol (r=−.03) or LDL cholesterol (r=.01) levels. Correlation coefficients were essentially similar between case patients and control subjects.

To examine the contribution of the small, dense LDL particles to the risk of IHD, a series of multivariate logistic analyses predicting the case-control status were performed (Table 2). As illustrated in model 1, the presence of small LDL particles (LDL-PPD≤25.64 nm) was associated with a 3.6-fold increase in the risk of IHD (95% CI, 1.5 to 8.8) independent of the potential confounding effects of diabetes, medication use, family history of IHD, and systolic blood pressure. Further adjustment for concomitant variations in plasma LDL cholesterol (model 2), TGs (model 3), and HDL cholesterol (model 4) had very little impact on the relationship between small LDL particles and the risk of IHD. Multivariate adjustment for apo B levels (model 5) and for the total/HDL cholesterol ratio (model 6) attenuated to some extent the association between LDL-PPD and the risk of IHD, but the impact of having small LDL particles on IHD risk remained significant (2.5-fold increase in the risk of IHD; 95% CI, 1.0 to 6.6). Although no longer significant from a statistical standpoint (P=.08), the risk associated with the presence of small, dense LDL particles was not lessened when the contributions of TGs, HDL cholesterol, and apo B to the risk of IHD were simultaneously taken into account (model 7). It is also important to point out that control for the contribution of LDL-PPD did not eliminate the association between apo B, the total/HDL cholesterol ratio, and the risk of IHD, suggesting that these variables and a high proportion of small, dense LDL particles may have some impact on the risk of IHD through independent mechanisms. These multivariate analyses were also performed while tertiles of plasma lipoprotein, lipid, and apolipoprotein levels were adjusted for. Similar results were obtained, because LDL particle size remained an independent predictor of the risk of IHD after adjustment for the contribution of plasma lipid and lipoprotein levels to IHD risk when investigated as tertiles. Exclusion of diabetic individuals did not affect the relationship between the dense LDL phenotype and the risk of IHD (not shown).

Additional analyses were performed after 25 pairs of individuals were eliminated in which either or both the case and the control subjects were using β-blockers or diuretics on a regular basis at the baseline evaluation. As shown in Table 3, the risk associated with the presence of small LDL particles (LDL-PPD≤25.64 nm) in this subsample of men not using medication remained elevated (OR, 5.1; P<.005) and was only slightly attenuated after the individual or simultaneous contributions of plasma TGs and apo B concentrations and of the total/HDL cholesterol ratio to IHD risk were controlled for.

Fig 3 presents the synergy between plasma apo B levels, the total/HDL cholesterol ratio, and small, dense LDL particles on the risk of IHD. Individuals with small LDL particles in the absence of elevated apo B concentrations (apo B levels <120 mg/dL, the median value of the distribution) were not at increased risk for IHD (OR, 1.0) compared with men with larger LDL particles and with relatively low apo B levels. Elevated apo B concentrations among individuals with large LDL particles resulted in a twofold increase in IHD risk, which did not reach statistical significance (P=.14), although the number of men in this particular subgroup was limited (n=32). Among these four groups, individuals having both elevated apo B levels and small LDL particles showed the greatest increase in IHD risk (OR, 6.2; 95% CI, 2.2 to 17.4; P<.001). A similar association was observed between LDL-PPD and the total/HDL cholesterol ratio, because only men with small LDL particles and with an elevated total/HDL cholesterol ratio were at greater risk of IHD (OR, 4.9; 95% CI, 1.9 to 12.7) compared with men having both large LDL particles and a ratio <6. Terms representing potential multiplicative interactions between LDL-PPD and other lipoprotein-lipid variables were tested in the multivariate model, and no interaction reached statistical significance.

We also investigated the relationship between the LDL-PPD as a continuous variable and the risk of IHD. A decrease of 0.65 nm (1 SD) in LDL-PPD was associated with a 35% increase in the risk of IHD (OR, 1.35; 95% CI, 0.97 to 1.89; P=.08) after adjustment for the confounding effects of diabetes, medication use, family history of IHD, and systolic blood pressure. Among lipid, lipoprotein, and apolipoprotein variables, apo B came out as the best and only significant predictor of IHD risk in multivariate stepwise logistic analyses (P=.002). LDL-PPD as a continuous variable did not contribute to the risk of IHD after the contribution of apo B levels to IHD risk had been considered.

Discussion

To the best of our knowledge, the present study is the first to have investigated the relationship between LDL-PPD and the risk of IHD using a prospective design in a population-based study. Our results are in agreement with previous case-control studies that have uniformly reported an increased prevalence of small, dense LDL particles in patients with IHD.^101112131415 In the two largest case-control studies involving male subjects, patients with the dense LDL phenotype were characterized by a 3-fold¹⁵ and a 4.5-fold²⁶ increase in the risk of IHD compared with those having a higher proportion of large LDL particles. In our study, men with small LDL particles (LDL-PPD≤25.64 nm) at baseline were at greater risk for the subsequent development of IHD than were individuals with larger particles (3.6- and 5.1-fold increases in odds ratios, respectively, when medication users were included or excluded from the analyses), independent of nonlipid risk factors such as diabetes mellitus, family history of IHD, and elevated systolic blood pressure. The relationship of small LDL particles to IHD was also independent of age, body mass index, alcohol consumption, and smoking, the covariables used in the pairing procedure of case patients and control subjects. These results suggest that the prospective association between the LDL particle size and the risk of IHD appears to be essentially of the same magnitude as the relationship reported in previous case-control investigations.¹⁷

Mechanisms responsible for the atherogenicity of small, dense LDL particles remain largely unexplained.¹⁷²⁷ It has been suggested that part of the risk associated with smaller, denser LDL particles may be related to the close relationship between this phenotype and disturbances in plasma lipid levels predictive of an increased risk of IHD.^6111727 The previous case-control reports that have compared the predictive value of the dense LDL phenotype with that of selected lipid and lipoprotein levels have in general concluded that the association between LDL phenotype and the risk of IHD was not independent of the concomitant variations in plasma TG, HDL cholesterol, or LDL cholesterol levels.^10121415 Only one study reported a residual association between LDL particle size and risk of IHD after adjustment for plasma TG levels.¹³ In the present study, the dense LDL phenotype remained an independent predictor of IHD risk after control for variations in plasma TG, LDL cholesterol, and HDL cholesterol concentrations and for the total/HDL cholesterol ratio, particularly when patients using β-blockers or diuretics were eliminated. These results suggest that the temporal association between the dense LDL phenotype and the risk of IHD may not be affected by differences in plasma lipoprotein and lipid levels at baseline and that small, dense LDL particles may increase the risk of IHD through additional mechanisms other than the concomitant variations in other atherogenic lipoproteins. It has been proposed that dense LDL may have lower binding affinity for the hepatic LDL apo B/E receptor.²⁸ LDL subfractions of intermediate density, generally found at higher concentrations in normolipidemic individuals, bind with a higher affinity to the LDL receptor and are degraded at greater rates than the denser LDL subfractions.²⁸ Denser LDL particles are also more susceptible to oxidation.²⁹³⁰³¹ The atherogenicity of dense LDL particles may also be related to their capacity for binding to the intimal proteoglycans.³²

We have previously reported that the hyperapolipoprotein B dyslipidemia was more prevalent in men who eventually developed IHD.¹⁹ Apo B is secreted as VLDL by the liver and remains associated with the particle until its clearance from the circulation as IDL and LDL.³³ There is systematically only one apo B per particle secreted,³³ and most fasting apo B is found in the LDL fraction. For this reason, apo B concentration can be considered to be a crude marker of LDL particle number. In the present report, the risk associated with smaller LDL particles appeared to be independent of the concomitant variations in apo B concentrations and thereby of particle number (Tables 2 and 3). Elevated apo B levels were also associated with an increased risk, independent of the LDL particle size. However, the presence of small LDL particles combined with elevated apo B concentrations resulted in the greatest increase (sixfold) in the risk of IHD (Fig 3). We also found that apo B concentration was the best metabolic predictor of IHD risk in multivariate analyses both in the present case-control analyses and in the study of the whole cohort of the Que´bec Cardiovascular Study.³⁴ These results support the hypothesis put forward by Tornvall et al¹¹ that LDL particle number, in addition to LDL composition, may be important with respect to IHD. These results may also have rather important clinical implications. They suggest that prevention and treatment of IHD should also be focused on reducing the number of atherogenic particles rather than only altering particle size. Although favorable increases in LDL-PPD have been observed after an intensive diet and exercise training program,³⁵ the lack of change or trivial alterations in particle size after interventions such as lipid-lowering therapy,³⁶³⁷³⁸ diet,³⁹ or exercise training⁴⁰ may not necessarily be indicative of an inefficient treatment, particularly in situations in which plasma apo B levels are substantially reduced by such approaches.⁴¹ In this regard, a reduction in plasma apo B concentrations after a low-fat diet has been reported in subjects with the small, dense LDL phenotype.⁴²

Summary

These prospective results from the Que´bec Cardiovascular Study suggest that a significant proportion of the risk associated with the presence of small, dense LDL particles may be independent of the concomitant variations in plasma lipid concentrations. Although characterization of the LDL phenotype and the measurement of plasma lipoprotein, lipid, and apolipoprotein levels both appear to provide independent information of IHD risk, the combination of small LDL particles and elevated apo B levels represents the metabolic state most predictive of IHD.

Selected Abbreviations and Acronyms

†Dr Moorjani died October 1, 1995.

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Footnotes