This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Syvänne, M. Articles by Taskinen, M.-R. Search for Related Content PubMed PubMed Citation Articles by Syvänne, M. Articles by Taskinen, M.-R.

(Circulation. 1995;92:364-370.)
© 1995 American Heart Association, Inc.

Articles

HDLs Containing Apolipoproteins A-I and A-II (LpA-I:A-II) as Markers of Coronary Artery Disease in Men With Non–Insulin- Dependent Diabetes Mellitus

Mikko Syvänne, MD; Juhani Kahri, MD; Kari S. Virtanen, MD; Marja-Riitta Taskinen, MD

From the First (M.S., K.S.V.) and Third (J.K., M-R.T.) Departments of Medicine, Helsinki (Finland) University Central Hospital.

Correspondence and reprint requests to Dr Mikko Syvänne, First Department of Medicine, Helsinki University Central Hospital, Haartmaninkatu 4, FIN-00290 Helsinki, Finland.

	Abstract

Top Abstract Introduction Methods Results Discussion References

 
Background Abnormalities in HDL and an increased risk of coronary artery disease (CAD) coexist in non–insulin-dependent diabetes mellitus (NIDDM). HDLs can be separated by their apolipoprotein (apo) content into particles containing apoA-I but not apoA-II (LpA-I) and those containing both apoA-I and apoA-II (LpA-I:A-II). The LpA-I particles have been suggested to be more effective in conferring protection against CAD than the LpA-I:A-II particles. However, data are sparse, and no studies have defined the role of these two classes of particles in NIDDM.

Methods and Results LpA-I and LpA-I:A-II particles were quantified by a differential electroimmunoassay in four groups of men with similar age and body mass index (BMI) distributions. Group 1 consisted of 50 patients with NIDDM and angiographically verified CAD; group 2, 50 men with CAD but no diabetes; group 3, 50 men with NIDDM but no CAD; and group 4, 31 healthy men. Serum apoA-I and apoA-II concentrations were measured by immunoturbidimetry, and HDL₂ and HDL₃ were separated by ultracentrifugation. Concentrations of LpA-I:A-II particles in group 1 were 13.8%, 18.3%, and 26.9% lower than in groups 2 through 4, respectively. In a two-by-two factorial ANOVA, adjusted for age and BMI, the differences were significant for both CAD (P<.001) and NIDDM (P<.001), with no interaction between the factors. These results were confirmed by comparable differences in the serum concentrations of apoA-I and apoA-II. LpA-I particles were related to the presence or absence of CAD (P=.013), but the difference was lost in a multivariate analysis. A low HDL₃ cholesterol concentration characterized both CAD (P=.002) and NIDDM (P=.024). HDL₂ cholesterol differed significantly with regard to the presence of NIDDM (P=.033) but only borderline with respect to CAD (P=.073).

Conclusions ApoA-II–containing lipoproteins and HDL₃ cholesterol are powerful markers of CAD in men with NIDDM.

Key Words: diabetes mellitus • lipoproteins • risk factors

	Introduction

Top Abstract Introduction Methods Results Discussion References

 
A low plasma concentration of HDL cholesterol constitutes one of the characteristic lipoprotein abnormalities in non–insulin-dependent diabetes mellitus (NIDDM).^{1 2} The well-documented high risk of coronary artery disease (CAD) among NIDDM subjects appears to be partially explained by low HDL cholesterol levels.³ However, HDL is heterogenous with respect to hydrated density and apolipoprotein (apo) content. Traditionally, HDL has been separated into HDL₂ and HDL₃, and the available data have suggested that a low level of the HDL₂ subfraction is associated with CAD in NIDDM patients.^{4 5} Recently, techniques have become available to separate HDL into lipoproteins (Lp) containing apoA-I but not apoA-II (LpA-I) and those containing both apoA-I and apoA-II (LpA-I:A-II).⁶ It has been suggested that LpA-I might represent "the antiatherogenic fraction of HDL"⁷ and that LpA-I:A-II particles might not actively participate in reverse cholesterol transport.⁸ Virtually no data are available on the associations of these particle classes with CAD in NIDDM.

Cholesterol contributes only about 15% to the total HDL mass and therefore may not be the ideal measure of HDL levels.⁹ Furthermore, the concentrations of HDL cholesterol, apoA-I, and LpA-I each appear to be influenced by different genetic and environmental factors.¹⁰ Thus, the question of whether apoA-I or apoA-II might provide additional insights into cardiovascular risk assessment has been raised, but so far the data in nondiabetic populations have been inconsistent.^{11 12} For example, Kwiterovich and coworkers¹³ reported that apoA-I was a better indicator of premature CAD than HDL₂ or HDL₃ cholesterol, but in contrast, Stampfer et al¹⁴ did not find the A apolipoproteins more useful in predicting the risk of myocardial infarction than cholesterol concentrations in the HDL subfractions. Even less is known about the predictive value of HDL apolipoproteins in NIDDM.² Briones et al¹⁵ found no association between the levels of apoA-I and CAD in NIDDM patients. Rönnemaa and colleagues¹⁶ reported that apoA-I levels were lower in NIDDM men who had suffered a myocardial infarction compared with NIDDM men with no evidence of CAD. However, the predictive power of apoA-I was lost in a multivariate model.

In the present study, we compared the concentrations of HDL cholesterol, HDL₂ and HDL₃ cholesterols, apoA-I, apoA-II, LpA-I, and LpA-I:A-II in four groups of men who had angiographically verified CAD, NIDDM, both of these conditions, or neither of them. Our main objective was to investigate which of these HDL-related parameters has the highest power to predict the presence or absence of CAD in men with NIDDM.

	Methods

Top Abstract Introduction Methods Results Discussion References

 
Subjects
Between September 1989 and February 1992, all patients with a history of NIDDM undergoing elective coronary angiography at the Helsinki University Central Hospital were screened for this study. Subjects 65 years of age or younger who had at least 50% stenosis in one or more major coronary branches were eligible. Exclusion criteria were hepatic, thyroid, or renal dysfunction and the use of lipid-lowering medications. Fifty consecutive men with NIDDM and CAD were selected for the present study (group 1, DM+CAD+). Groups 2 through 4 were selected to have distributions of age and body mass index (BMI) similar to those of group 1 by enrolling approximately equal proportions of subjects in four age and BMI categories (young and lean, young and obese, old and lean, old and obese), with the median values of group 1 as cutoff points. The above-mentioned exclusion criteria also were applied to these groups, but the control groups were otherwise selected randomly, eg, without any regard to lipid levels. Group 2 (DM-CAD+) consisted of 50 men who also had angiographically verified CAD but no history of diabetes and normal fasting blood glucose and glycosylated hemoglobin A_1c (HbA_1c) levels. The 50 men in group 3 (DM+CAD–) were recruited among the NIDDM patients of outpatient diabetes clinics. They were required to have no history or symptoms of CAD. To exclude silent ischemic heart disease, the patients underwent a maximal exercise test with tomographic ²⁰¹Tl imaging¹⁷ with normal results. Group 4 (DM–CAD–) consisted of 31 apparently healthy men. They had no known diseases; were not taking any medication; and had normal cardiovascular physical examinations, resting ECGs, symptom-limited exercise ECGs, and laboratory screens, including fasting blood glucose and HbA_1c.

Table 1 lists selected characteristics of the study groups. By selection, the groups were similar with respect to age and BMI. The proportion of current smokers was similar in all groups. Reported alcohol intake tended to be slightly but nonsignificantly higher in the groups without CAD. As expected, fasting insulin levels were significantly higher in NIDDM than in nondiabetic groups, but hyperinsulinemia also appeared to characterize CAD patients, regardless of diabetes.

View this table: [in this window] [in a new window]   Table 1. Clinical and Biochemical Characteristics of the Study Groups

All CAD patients (groups 1 and 2) had angina pectoris. Their cardiac symptoms and findings were comparable. We found that 37 patients in group 1 (DM+CAD+) and 38 in group 2 (DM–CAD+) had angina in mild exertion (New York Heart Association [NYHA] class III); the rest were in NYHA class II. In addition, 28 patients in group 1 and 31 in group 2 had a history of myocardial infarction, but not within 3 months before participating in this study. No one had clinically or radiologically evident congestive heart failure at the time of the study. We also found that 10, 11, and 29 patients in group 1 and 7, 12, and 31 patients in group 2 had one-, two-, and three-vessel CAD, respectively. Comparable numbers of subjects in both groups subsequently underwent coronary bypass surgery (group 1, 33; group 2, 41) or percutaneous transluminal coronary angioplasty (12 and 8, respectively). Left ventricular ejection fractions were similar (group 1, 56±15%; group 2, 60±15%). The groups were also similar with respect to exercise capacity. The majority of patients in both groups were taking ß-blockers (group 1, 41 subjects; group 2, 40 subjects), long-acting nitrates (43 and 47, respectively), and aspirin (47 and 46, respectively). Digoxin, diuretics, and angiotensin-converting enzyme (ACE) inhibitors were used by 7, 8, and 6 patients in group 1 and by 3, 6, and 4 patients in group 2, respectively. The only significant difference in medication between the two groups was the use of calcium channel blockers: 41 in group 1, 30 in group 2 (P<.03).

All diabetic subjects (group 1, DM+CAD+, and group 3, DM+CAD–) fulfilled the World Health Organization (WHO) definition of NIDDM.¹⁸ In both groups, 12 patients were treated by diet only. Thirty subjects in group 1 and 32 in group 3 were taking oral hypoglycemic agents (mainly sulfonylureas). Eight patients in group 1 and 6 in group 3 were on insulin, either alone or in combination with a sulfonylurea. Because of the selection criteria, no one had gross proteinuria, and only 8 subjects in group 1 and 4 in group 3 had microalbuminuria (albumin excretion rate >30 µg/min but <200 µg/min). Hypertension was more prevalent in group 1 than in group 3 (Table 1). In group 3, 6 patients were taking ß-blockers, 3 were taking calcium channel blockers, 7 were taking ACE inhibitors, and no one was taking diuretics. Glycemic control was moderate in both groups (Table 1).

All subjects gave their informed consent to participate in the study. The protocol was approved by the Ethical Committee of the First Department of Medicine, University of Helsinki.

Diagnostic Procedures and Laboratory Analyses
Blood samples were obtained in the morning after an overnight fast. Thereafter, heparin (100 U/ kg of body weight; maximum dose 10 000 U) was given as an intravenous bolus injection. Postheparin samples were taken 5 and 15 minutes later, and lipoprotein lipase (LPL) and hepatic lipase (HL) activities were determined as described.¹⁹ Lipoprotein fractions (VLDL, d<1.006; IDL, d=1.006 to 1.019; LDL, d 1.019 to 1.063; HDL; HDL₂, d=1.063 to 1.125; and HDL₃, d=1.125 to 1.210) were separated from fresh fasting sera by sequential flotation in an ultracentrifuge as previously described in detail.²⁰ Cholesterol and triglyceride (TG) levels were measured in serum and in the lipoprotein fractions by automated enzymatic methods.²¹ Serum apoA-I and apoA-II concentrations were measured by an immunoturbidimetric method with commercially available kits (Boehringer Mannheim); apoB was measured by an immunochemical assay (Orion Diagnostica). The concentration of LpA-I particles was quantified by use of a differential electroimmunoassay with hydrated agarose gels containing monospecific antibodies against apoA-I and A-II (Sebia) as described elsewhere in detail.^{22 23} This method directly quantifies the LpA-I particles as the concentration of apoA-I in these particles. The concentration of LpA-I:A-II particles was calculated by subtracting the concentration of LpA-I particles from the turbidimetrically measured total concentration of apoA-I in serum. In our laboratory, the interassay variations for apoA-I, apoA-II, and LpA-I particle concentrations are 3.5%, 2.1%, and 7.3%, respectively.²³ ApoE phenotyping was performed in serum.²⁴

Blood glucose, serum-free insulin, C-peptide, and HbA_1c were determined as described.²⁵ Standardized interviews were performed to acquire information about the subjects' smoking and drinking habits. BMI was calculated as weight in kilograms divided by height in meters squared. The waist-to-hip ratio was calculated as the smallest girth between the lowest ribs and the umbilicus divided by the largest horizontal girth between waist and thigh. The CAD patients' angiographic and other relevant data were obtained from their cardiologists' reports.

Statistical Analyses
Data are expressed as mean±SD or as mean and 95% CI of the mean. Bartlett's test was used to evaluate the homogeneity of variances. Logarithmic transformations were performed, if necessary, to achieve a normal distribution or to stabilize variances across the study groups. The data for the transformed variables are expressed as geometric mean and 95% CI. The data with continuous distributions were analyzed by an ANOVA in a two-by-two factorial design²⁶ ; the factors were the presence or absence of CAD and NIDDM. The possibility of an interaction between the two factors was also tested. If there is no evidence of a significant interaction, then a difference with respect to one factor (eg, CAD) is independent of the other (NIDDM). The analyses were adjusted for age and BMI by including the enrollment category (eg, young and lean) as a third factor in the ANOVA model. Using age and BMI as continuous covariates yielded essentially similar results (data not shown). Differences in proportions were assessed by the ² test with continuity correction. Pearson's coefficients were calculated to study correlations.

The effect of potential confounders was accounted for by entering them into the ANOVA model as covariates. The following covariates were used: VLDL TG concentration, LPL and HL activities, waist-to-hip ratio, fasting serum insulin concentration, glycemic control (HbA_1c), alcohol use, hypertension, and the use of ß-blocking drugs.

From a clinical viewpoint, it is of interest if new laboratory assays, such as LpA-I and LpA-I:A-II particles, are more powerful in identifying individuals at risk of CAD than those already in common use. Therefore, we constructed six models with various combinations of the HDL-related parameters and tested their power to predict the group allocation of the subjects by logistic regression and the likelihood ratio test.²⁷ In one set of analyses, the dependent variable was the presence or absence of CAD; in another, the presence or absence of NIDDM. The first model included serum TG, serum cholesterol, and HDL cholesterol, which are readily available in everyday clinical practice. To this battery of tests we added HDL₂ and HDL₃ cholesterol (model 2), apoA-I and apoA-II (model 3), or LpA-I and LpA-I:A-II (model 4) and compared each of these three new models with model 1. Model 5 consisted of the variables in model 2 plus apoA-I and apoA-II. The purpose of comparing models 5 and 2 was to see whether measuring serum apolipoprotein concentrations improved predictive accuracy compared with the more traditional HDL subfractionation by ultracentrifugation. Finally, model 6 contained all the variables in model 5 plus LpA-I and LpA-I:A-II levels. Models 5 and 6 were compared to evaluate the contribution of apolipoprotein-specific HDL particles when all other variables had been taken into account.

	Results

Top Abstract Introduction Methods Results Discussion References

 
Lipoprotein and Apolipoprotein Concentrations
Overall, there were no significant interactions between the factors of CAD and NIDDM (Table 2). Group 1 (DM+CAD+) had the highest serum and VLDL TG concentrations, which were significantly associated with the presence of CAD (Table 2). LDL cholesterol was not related to CAD, but the NIDDM groups had significantly lower concentrations than the nondiabetic subjects. Group 1 had the lowest HDL and HDL₂ cholesterol levels, but this appeared to be related to NIDDM rather than to CAD (Table 2). In contrast, a low HDL₃ cholesterol concentration was an efficient marker of both NIDDM and CAD. Group 1 had 7.5%, 8.4%, and 18.9% lower HDL₃ cholesterol levels than groups 2 (DM–CAD+), 3 (DM+CAD–), and 4 (DM–CAD–), respectively.

View this table: [in this window] [in a new window]   Table 2. Lipoprotein-Related Variables in the Study Groups

Both of the major HDL apolipoproteins were markedly lower in group 1 than in the other three groups; the contrasts between CAD and non-CAD and those between NIDDM and nondiabetic groups were highly significant (Table 2). ApoA-I concentrations were 9.8%, 15.1%, and 21.1% and apoA-II levels were 16.4%, 20.7%, and 25.9% lower in group 1 than in groups 2 through 4, respectively.

The postheparin plasma LPL activity tended to be lowest and HL activity highest in the DM+CAD+ group, but the differences did not reach statistical significance. The apoE phenotype distributions were essentially similar within all groups (Table 2).

LpA-I and LpA-I:A-II Particles
The apolipoprotein-specific HDL particles differed markedly in their power to predict the presence of CAD or NIDDM in the study population. Although group 1 had the lowest concentration of LpA-I particles, they were only 1.0%, 7.6%, and 8.3% lower than in groups 2 through 4, respectively. Nevertheless, the contrast between CAD and non-CAD was statistically significant (P=.013; Table 2). The presence or absence of NIDDM did not appear to influence LpA-I concentrations.

LpA-I:A-II particles, measured as their apoA-I concentration, were 13.8%, 18.3%, and 26.9% lower in group 1 than in groups 2 through 4, respectively. The presence of both CAD (P<.001) and NIDDM (P<.001) was related to the concentration of LpA-I:A-II particles, and there was no evidence of an interaction between CAD and NIDDM (P=.946); ie, a low level of these particles was independently related to both diseases.

As Table 3 shows, the concentration of LpA-I particles was in close correlation with serum apoA-I and HDL₂ cholesterol concentrations. LpA-I:A-II particles were related to both apoA-I and A-II levels and more closely related to HDL₃ than to HDL₂ cholesterol. There was a moderate positive relation between the postheparin plasma LPL activity and both types of HDL particles, especially LpA-I:A-II. These correlations were essentially similar when calculated within each of the study groups (data not shown).

View this table: [in this window] [in a new window]   Table 3. Pearson's Correlation Coefficients Between Selected HDL Variables and Triglyceride Lipases

Analysis of Covariance
Several biochemical variables and clinical characteristics that may influence HDL metabolism and thus are potential confounders in between-group comparisons were evaluated by an ANCOVA in a similar two-by-two factorial design as the unadjusted analyses. Table 4 shows that HDL₃ cholesterol, apoA-I, apoA-II, and LpA-I:A-II were strongly related to the presence of CAD, even when the covariates were taken into account. However, HDL cholesterol, HDL₂ cholesterol, and LpA-I did not significantly differ with respect to CAD in the adjusted analysis.

View this table: [in this window] [in a new window]   Table 4. Two-By-Two Factorial ANCOVA

Comparison of HDL-Related Parameters as Markers of CAD
The logistic regression model containing serum TG, cholesterol, and HDL cholesterol concentrations was unable to predict the presence of CAD in this population (Table 5). Adding HDL subfractions separated by ultracentrifugation (HDL₂ and HDL₃), apoA-I and apoA-II, or LpA-I and LpA-I:A-II concentrations each greatly enhanced the power of the model. Model 5 was superior to model 2 (Table 5), showing that the HDL apolipoproteins improved the model even when HDL₂ and HDL₃ were already taken into account. Finally, a model containing LpA-I and LpA-I:A-II in addition to all the other variables was slightly but significantly more powerful than the model without the apolipoprotein-specific HDL subfractions.

View this table: [in this window] [in a new window]   Table 5. Likelihood Ratio Tests Comparing the Power of Various Combinations of HDL-Related Parameters to Predict the Presence or Absence of Coronary Artery Disease and Non–Insulin-Dependent Diabetes Mellitus

A similar analysis to predict the presence or absence of NIDDM showed, in agreement with the data in Table 2, that only models 4 and 6 (containing LpA-I and LpA-I:A-II) improved the predictive power of measuring only the commonly used lipid variables of model 1.

	Discussion

Top Abstract Introduction Methods Results Discussion References

 
The first experimental studies with HDL particles separated by their apolipoprotein content suggested that LpA-I promotes the efflux of cholesterol from cells, while LpA-I:A-II particles may actually antagonize this effect.^{28 29} However, subsequent experiments with different cellular models indicated that both LpA-I and LpA-I:A-II may play a role in reverse cholesterol transport.^{30 31} According to a recent review, the role of each of the HDL apolipoproteins in the regulation of cholesterol metabolism and movement remains to be clarified.³² In one cross-sectional clinical study, LpA-I levels were lower in CAD patients than in control subjects, whereas there was no difference in LpA-I:A-II particles.⁷ So far, very few studies on this topic have been published, but evidence is emerging that a low concentration of LpA-I:A-II particles may also characterize CAD patients^{33 34 35 36} or subjects at great risk of CAD³⁷ compared with healthy control subjects.

The results of the present study support the concept that LpA-I:A-II particles may exert a protective effect against CAD in the population studied, men with NIDDM. The concentrations of these particles were low in the DM+CAD+ group, and statistical analyses suggested that this was significantly related to the presence of both CAD and NIDDM. Methodological considerations give further weight to the results because serum apoA-I and apoA-II concentrations were also clearly lowest in patients with NIDDM and CAD. Because LpA-I:A-II particles were measured on the basis of their apoA-I content, two independent assays are in agreement.

Like several previous investigators,^{7 33 34 35 36} we also found lower LpA-I levels in the CAD groups compared with the CAD-free control groups, but this difference was much smaller than that observed in LpA-I:A-II concentrations and did not remain significant after adjustment for potential confounders in ANCOVA.

However, the apolipoprotein-specific HDL particles contributed relatively little to the power to predict the presence of CAD when the serum concentrations of apoA-I and A-II were known (Table 5, model 6 versus model 5). This may be due in part to the smaller analytical variation in the apoA-II assay compared with the indirect quantification of LpA-I:A-II particles by an electroimmunoassay. Thus, the determination of apoA-II should be a useful surrogate for measuring LpA-I:A-II particles because most serum apoA-II resides in these lipoproteins, although a minor lipoprotein species may contain only apoA-II.³⁸ Serum apoA-I concentrations were also low in the DM+CAD+ group. This seems to reflect primarily the paucity of LpA-I:A-II particles because the between-group differences in LpA-I levels were small.

When considering the mechanisms underlying the low LpA-I:A-II levels in patients with NIDDM and CAD, we must take into account the interactions with other lipoprotein classes, especially the TG-rich VLDL and chylomicrons. Hypertriglyceridemia results in enrichment of HDL₂ with TG, and hydrolysis of the excess TG by HL converts these particles into smaller, denser HDL₃.³⁹ LpA-I:A-II particles are the preferred substrate for HL compared with LpA-I.⁴⁰ Therefore, low levels of both LpA-I:A-II and HDL₃ cholesterol in our DM+CAD+ patients with high HL activities are in good agreement with these experimental studies. However, our ANCOVA indicated that the between-group differences observed in LpA-I:A-II levels remained significant, even when VLDL TG level and HL activity and characteristics such as alcohol consumption and the use of ß-blockers were taken into account. Thus the data indirectly suggest that genetic factors¹⁰ may be important determinants of LpA-I:A-II concentrations.

NIDDM, regardless of CAD, was also associated with low concentrations of LpA-I:A-II and apoA-II (Table 2). The mechanisms underlying the low levels of apoA-II–containing lipoproteins in NIDDM are currently unclear and should be investigated in future studies.

Although a low HDL cholesterol was significantly (P=.016) related to NIDDM, the factorial ANOVA indicated only a borderline-significant (P=.044) relation to CAD in the present study, in agreement with recent prospective data by Uusitupa and coworkers.⁴¹ Of the classic HDL subfractions, we found that HDL₃ cholesterol was a much stronger predictor than HDL₂ cholesterol. Our finding of low HDL₃ levels in CAD patients adds to previous similar data in nondiabetic populations, as reviewed by Miller.¹¹ More recently, Stampfer et al¹⁴ found in a prospective study that both HDL₂ and HDL₃ cholesterols were predictors of myocardial infarction. Our data are internally consistent because we found a closer correlation between LpA-I:A-II and HDL₃ than HDL₂ cholesterol (Table 3) and are in agreement with earlier data that indicate that LpA-I:A-II particles reside primarily (but not entirely) in the HDL₃ size range.⁴²

So far, only limited data exist on HDL subfractions in patients with NIDDM and CAD. Our results regarding HDL₂ and HDL₃ contrast with cross-sectional⁴ and prospective⁵ data by Laakso et al. In these studies, HDL₂ cholesterol was a marker or a predictor of clinical CAD events, whereas HDL₃ cholesterol was not. The reasons for this discrepancy remain speculative. One possibility is that our DM+CAD+ patients, who had had a clinical indication of coronary angiography and most of whom required a revascularization procedure, represented more severe CAD than the population-based sample of Laakso and coworkers. It is also possible that CAD patients who present primarily with severe angina pectoris, like our study group, differ from those who experience a myocardial infarction or a cardiac death.

As always in a cross-sectional study, caution must be exercised in deducing causality. In our study, differences in alcohol consumption and the use of ß-blockers are obvious potential confounders that are known to influence HDL metabolism. An ANCOVA did not suggest that these factors were responsible for the differences observed in LpA-I:A-II particle concentrations. However, it is an inherent limitation of cross-sectional case-control studies that unrecognized confounders or selection bias may distort the results. Our results imply that the measurement of apoA-II-containing lipoproteins and HDL subfraction cholesterol concentrations is imperative in future prospective studies of CAD risk factors, especially in NIDDM patients.

CAD in our patients was diagnosed by coronary angiography but excluded in the control groups by noninvasive testing. Thus, the subjects in groups 3 and 4 may have had mild subclinical coronary atherosclerosis. However, even a normal-appearing angiogram does not guarantee the absence of incipient atherosclerosis.^{43 44} Moreover, our control groups consisted of clinically healthy individuals, whereas control subjects selected on the basis of a normal arteriogram would have had a clinical indication to undergo the examination. In any case, the putative subclinical atherosclerosis in our control groups does not invalidate the results because it would tend, if anything, to attenuate the between-group differences.

Recently, Rastogi and coworkers⁴⁵ published preliminary data from their ongoing study of HDL particles in diabetes. Their data are in perfect agreement with ours in that low levels of LpA-I:A-II particles characterize men with NIDDM and CAD compared with healthy subjects. As in our study, they also had a group of diabetic patients without CAD. However, this control group was small and heterogenous with respect to the type of diabetes. Thus, the crucial finding in our study, that low LpA-I:A-II concentrations are a characteristic of CAD in NIDDM patients and not merely a marker of diabetes as such, awaits confirmation in future studies.

To conclude, in this cross-sectional study, low concentrations of LpA-I:A-II particles, serum apoA-I and apoA-II concentrations, and HDL₃ cholesterol were powerful indicators of CAD in men with NIDDM and among the nondiabetic men in this study population. Highly significant differences in these variables existed between a group with NIDDM and angiographically assessed CAD (DM+CAD+) and a control group in which CAD had been ruled out by noninvasive testing, whether the control subjects had diabetes (DM+CAD–) or not (DM–CAD–). Serum apoA-II concentration may be a powerful risk marker of CAD in NIDDM patients, given the high analytical precision and cheaper price of the assay compared with the quantification of LpA-I:A-II particles. Taken together with previously published data, our results suggest that different patient groups may have variable perturbations in their HDL metabolism that expose them to the development of atherosclerosis.

	Acknowledgments

 
This work was supported by research grants from the Finnish Cardiac Research Foundation, the Aarne Koskelo Foundation, and the Sigrid Juselius Foundation. We gratefully acknowledge the excellent laboratory work by Hannele Hilden, Leena Lehikoinen, Ritva Marjanen, Anne-Mari Pylkkänen, Sirpa Rannikko, and Sirkka-Liisa Runeberg. Assistance from Aino Korpela, RN, Anita Leppämäki, RN, Tina Svahn, RN, and several other people was vital to patient recruitment. We thank Juni Palmgren, PhD, and Mikko Virtanen, MSc, for statistical advice.

Received December 27, 1994; accepted January 17, 1995.

	References

Top Abstract Introduction Methods Results Discussion References

 
1. Howard BV. Lipoprotein metabolism in diabetes mellitus. J Lipid Res. 1987;28:613-628.[Medline] [Order article via Infotrieve]

2. Taskinen M-R. Hyperlipidaemia in diabetes. In: Betteridge DJ, ed. Baillière's Clinical Endocrinology and Metabolism. London, UK: Baillière, Tindall; 1990;4:743-775.

3. Pyörälä K, Laakso M, Uusitupa M. Diabetes and atherosclerosis: an epidemiologic view. Diabetes Metab Rev. 1987;3:463-524. [Medline] [Order article via Infotrieve]

4. Laakso M, Voutilainen E, Pyörälä K, Sarlund H. Association of low HDL and HDL₂ cholesterol with coronary heart disease in non–insulin-dependent diabetics. Arteriosclerosis. 1985;5:653-658. [Abstract/Free Full Text]

5. Laakso M, Lehto S, Penttilä I, Pyörälä K. Lipids and lipoproteins predicting coronary heart disease mortality and morbidity in patients with non–insulin-dependent diabetes. Circulation. 1993;88:1421-1430. [Abstract/Free Full Text]

6. Cheung MC, Albers JJ. Characterization of lipoprotein particles isolated by immunoaffinity chromatography: particles containing A-I and A-II and particles containing A-I but no A-II. J Biol Chem. 1984;259:12201-12209. [Abstract/Free Full Text]

7. Puchois P, Kandoussi A, Fievet P, Fourrier JL, Bertrand M, Koren E, Fruchart J-C. Apolipoprotein A-I containing lipoproteins in coronary artery disease. Atherosclerosis. 1987;68:35-40. [Medline] [Order article via Infotrieve]

8. Rader DJ, Castro GC, Zech LA, Fruchart J-C, Brewer HJ Jr. In vivo metabolism of apolipoprotein A-I on high density lipoprotein particles LpA-I and LpA-I,A-II. J Lipid Res. 1991;32:1849-1859. [Abstract]

9. Eisenberg S. High density lipoprotein metabolism. J Lipid Res. 1984;25:1017-1058. [Medline] [Order article via Infotrieve]

10. Steinmetz J, Boerwinkle E, Gueguen R, Visvikis S, Henny J, Siest G. Multivariate genetic analysis of high density lipoprotein particles. Atherosclerosis. 1992;92:219-227. [Medline] [Order article via Infotrieve]

11. Miller NE. Associations of high-density lipoprotein subclasses and apolipoproteins with ischemic heart disease and coronary atherosclerosis. Am Heart J. 1987;113:589-597. [Medline] [Order article via Infotrieve]

12. Hamsten A. Apolipoproteins, dyslipoproteinaemia and premature coronary heart disease. Acta Med Scand. 1988;223:389-403. [Medline] [Order article via Infotrieve]

13. Kwiterovich PO Jr, Coresh J, Smith HH, Bachorik PS, Derby CA, Pearson TA. Comparisons of the plasma levels of apolipoproteins B and A-1, and other risk factors in men and women with premature coronary artery disease. Am J Cardiol. 1992;69:1015-1021. [Medline] [Order article via Infotrieve]

14. Stampfer MJ, Sacks FM, Salvini S, Willett WC, Hennekens CH. A prospective study of cholesterol, apolipoproteins, and the risk of myocardial infarction. N Engl J Med. 1991;325:378-381.

15. Briones ER, Mao SJT, Palumbo P, O'Fallon WM, Chenoweth W, Kottke BA. Analysis of plasma lipids and apolipoproteins in insulin-dependent and noninsulin-dependent diabetics. Metabolism. 1984;33:42-49. [Medline] [Order article via Infotrieve]

16. Rönnemaa T, Laakso M, Kallio V, Pyörälä K, Marniemi J, Puukka P. Serum lipids, lipoproteins, and apolipoproteins and the excessive occurrence of coronary heart disease in non–insulin-dependent diabetic patients. Am J Epidemiol. 1989;130:632-645. [Abstract/Free Full Text]

17. Kupari M, Virtanen KS, Turto H, Viitasalo M, Mänttäri M, Lindroos M, Koskela E, Leinonen H, Pohjola-Sintonen S, Heikkilä J. Exclusion of coronary artery disease by exercise thallium-201 tomography in patients with aortic valve stenosis. Am J Cardiol. 1992;70:635-640. [Medline] [Order article via Infotrieve]

18. WHO Expert Committee on Diabetes Mellitus. Second report. Geneva: World Health Organization; 1980. World Health Organization technical report series 646.

19. Huttunen JK, Ehnholm C, Kinnunen PJ, Nikkilä EA. An immunochemical method for selective measurement of two triglyceride lipases in human postheparin plasma. Clin Chim Acta. 1975;63:335-347. [Medline] [Order article via Infotrieve]

20. Taskinen M-R, Kuusi T, Helve E, Nikkilä E, Yki-Järvinen H. Insulin therapy induces antiatherogenic changes of serum lipoproteins in non–insulin-dependent diabetes. Arteriosclerosis. 1988;8:168-177. [Abstract/Free Full Text]

21. Syvänne M, Vuorinen-Markkola H, Hilden H, Taskinen M-R. Gemfibrozil reduces postprandial lipemia in non–insulin-dependent diabetes mellitus. Arterioscler Thromb. 1993;13:286-295. [Abstract/Free Full Text]

22. Parra HJ, Mezdour H, Ghalim N, Bard JM, Fruchart J-C. Differential electroimmunoassay of human Lp A-I lipoprotein particles on ready-to-use plates. Clin Chem. 1990;36:1431-1435. [Abstract/Free Full Text]

23. Kahri J, Vuorinen-Markkola H, Tilly-Kiesi M, Lahdenperä S, Taskinen M-R. Effect of gemfibrozil on high density lipoprotein subspecies in non-insulin dependent diabetes mellitus: relation to lipolytic enzymes and to the cholesteryl ester transfer protein activity. Atherosclerosis. 1993;102:79-89. [Medline] [Order article via Infotrieve]

24. Havekes L, de Knijeff P, Beisiegel U, Havinga J, Smith M, Klasen E. A rapid micromethod for apolipoprotein E phenotyping directly in serum. J Lipid Res. 1987;28:455-463. [Abstract]

25. Vuorinen-Markkola H, Yki-Järvinen H, Taskinen M-R. Lowering of triglycerides by gemfibrozil affects neither the glucoregulatory nor antilipolytic effect of insulin in noninsulin-dependent diabetic patients. Diabetologia. 1993;36:161-169. [Medline] [Order article via Infotrieve]

26. SYSTAT Inc. SYSTAT for Windows. Statistics, Version 5 Edition. Evanston, Ill: SYSTAT Inc; 1992:210-379.

27. Hosmer DW, Lemeshow S. Applied Logistic Regression. New York, NY: John Wiley & Sons, Inc; 1989.

28. Barbaras R, Puchois P, Fruchart J-C, Ailhaud G. Cholesterol efflux from cultured adipose cells if mediated by LpA_I particles but not by LpA_I:A_II particles. Biochem Biophys Res Commun. 1987;142:63-69. [Medline] [Order article via Infotrieve]

29. Barkia A, Puchois P, Ghalim N, Torpier G, Barbaras R, Ailhaud G, Fruchart J-C. Differential role of apolipoprotein A_I-containing particles in cholesterol efflux from adipose cells. Atherosclerosis. 1991;87:135-146. [Medline] [Order article via Infotrieve]

30. Johnson WJ, Kilsdonk EPC, van Tol A, Phillips MC, Rothblat GH. Cholesterol efflux from cells to immunopurified subfractions of human high density lipoprotein: LP-AI and LP-AI/AII. J Lipid Res. 1991;32:1993-2000. [Abstract]

31. Oikawa S, Mendez AJ, Oram JF, Bierman EL, Cheung MC. Effects of high-density lipoprotein particles containing apo A-I, with or without apo A-II, on intracellular cholesterol efflux. Biochim Biophys Acta. 1993;1165:327-334.[Medline] [Order article via Infotrieve]

32. Johnson WJ, Mahlberg FH, Rothblat GH, Phillips MC. Cholesterol transport between cells and high-density lipoproteins. Biochim Biophys Acta. 1991;1085:273-298. [Medline] [Order article via Infotrieve]

33. Fruchart JC, Ailhaud G. Recent progress in the study of apoA-containing lipoprotein particles. Prog Lipid Res. 1991;30:145-150. [Medline] [Order article via Infotrieve]

34. Parra HJ, Arveiler D, Evans AE, Cambou JP, Amouyel P, Bingham A, McMaster D, Schaffer P, Douste-Blazy P, Luc G, Richard JL, Ducimetière P, Fruchart JC, Cambien F. A case-control study of lipoprotein particles in two populations at contrasting risk for coronary heart disease: the ECTIM Study. Arterioscler Thromb. 1992;12:701-707. [Abstract/Free Full Text]

35. Genest JJ, Bard JM, Fruchart J-C, Ordovas JM, Wilson PFW, Schaefer EJ. Plasma apolipoprotein A-I, A-II, B, E and CIII-containing particles in men with premature coronary artery disease. Atherosclerosis. 1991;90:149-157. [Medline] [Order article via Infotrieve]

36. Coste-Burel M, Mainard F, Chivot L, Auget JL, Madec Y. Study of lipoprotein particles LpAI and LpAI:AII in patients before coronary bypass surgery. Clin Chem. 1990;36:1889-1891. [Abstract]

37. Cachera C, Kandoussi A, Equagoo K, Fruchart J-C, Tacquet A. Evaluation of apolipoprotein A-I containing particles in chronic renal failure patients undergoing hemodialysis. Am J Nephrol. 1990;10:171-172. [Medline] [Order article via Infotrieve]

38. Bekaert ED, Alaupovic P, Knight-Gibson C, Ayrault-Jarrier M. The lipoprotein A-II (LP-A-II) family in normolipidemic subjects and patients with apoA-I deficiency. Prog Lipid Res. 1991;30:151-157. [Medline] [Order article via Infotrieve]

39. Patsch JR, Prasad S, Gotto AM Jr, Bengtsson-Olivecrona G. Postprandial lipemia: a key for the conversion of high density lipoprotein₂ into high density lipoprotein₃ by hepatic lipase. J Clin Invest. 1984;74:2017-2023.

40. Mowri H-O, Patsch W, Smith LC, Gotto AM Jr, Patsch JR. Different reactivities of high density lipoprotein₂ subfractions with hepatic lipase. J Lipid Res. 1992;33:1269-1279. [Abstract]

41. Uusitupa MIJ, Niskanen LK, Siitonen O, Voutilainen E, Pyörälä K. Ten-year cardiovascular mortality in relation to risk factors and abnormalities in lipoprotein composition in type 2 (non–insulin-dependent) diabetic and non-diabetic subjects. Diabetologia. 1993;36:1175-1184. [Medline] [Order article via Infotrieve]

42. Cheung MC, Segrest JP, Albers JJ, Cone JT, Brouillette CG, Chung BH, Kashyap M, Glasscock MA, Anantharamaiah GM. Characterization of high density lipoprotein subspecies: structural studies by single vertical spin ultracentrifugation and immunoaffinity chromatography. J Lipid Res. 1987;28:913-929. [Abstract]

43. McPherson DD, Hiratzka LF, Lamberth WC, Brandt B, Hunt M, Kieso RA, Marcus ML, Kerber RE. Delineation of the extent of coronary atherosclerosis by high frequency epicardial echocardiography. N Engl J Med. 1987;316:304-309. [Abstract]

44. Glagov S, Weisenberg E, Zarins CK, Stankunavicius R, Kolettis GJ. Compensatory enlargement of human atherosclerotic coronary arteries. N Engl J Med. 1987;316:1371-1375. [Abstract]

45. Rastogi A, Bren ND, Hallaway BJ, Kottke BA. Immunomagnetic separation of subpopulations of apolipoprotein A-I. Mayo Clin Proc. 1994;69:137-143.[Medline] [Order article via Infotrieve]

This article has been cited by other articles:

				D. C. Chan, G. F. Watts, E. M.M. Ooi, K.-A. Rye, J. Ji, A. G. Johnson, and P. H. R. Barrett Regulatory Effects of Fenofibrate and Atorvastatin on Lipoprotein A-I and Lipoprotein A-I:A-II Kinetics in the Metabolic Syndrome Diabetes Care, November 1, 2009; 32(11): 2111 - 2113. [Abstract] [Full Text] [PDF]

				C. Wolfrum, J. J. Howell, E. Ndungo, and M. Stoffel Foxa2 Activity Increases Plasma High Density Lipoprotein Levels by Regulating Apolipoprotein M J. Biol. Chem., June 13, 2008; 283(24): 16940 - 16949. [Abstract] [Full Text] [PDF]

				E. M. M. Ooi, G. F. Watts, P. J. Nestel, D. Sviridov, A. Hoang, and P. H. R. Barrett Dose-Dependent Regulation of High-Density Lipoprotein Metabolism with Rosuvastatin in the Metabolic Syndrome J. Clin. Endocrinol. Metab., February 1, 2008; 93(2): 430 - 437. [Abstract] [Full Text] [PDF]

				R. S. Birjmohun, G. M. Dallinga-Thie, J. A. Kuivenhoven, E. S.G. Stroes, J. D. Otvos, N. J. Wareham, R. Luben, J. J.P. Kastelein, K.-T. Khaw, and S. M. Boekholdt Apolipoprotein A-II Is Inversely Associated With Risk of Future Coronary Artery Disease Circulation, October 30, 2007; 116(18): 2029 - 2035. [Abstract] [Full Text] [PDF]

				J. Ji, G. F. Watts, A. G. Johnson, D. C. Chan, E. M. M. Ooi, K.-A. Rye, A. P. Serone, and P. H. R. Barrett High-Density Lipoprotein (HDL) Transport in the Metabolic Syndrome: Application of a New Model for HDL Particle Kinetics J. Clin. Endocrinol. Metab., March 1, 2006; 91(3): 973 - 979. [Abstract] [Full Text] [PDF]

				G. Luc, J.-M. Bard, J. Ferrieres, A. Evans, P. Amouyel, D. Arveiler, J.-C. Fruchart, P. Ducimetiere, and on behalf of the PRIME Study Group Value of HDL Cholesterol, Apolipoprotein A-I, Lipoprotein A-I, and Lipoprotein A-I/A-II in Prediction of Coronary Heart Disease: The PRIME Study Arterioscler Thromb Vasc Biol, July 1, 2002; 22(7): 1155 - 1161. [Abstract] [Full Text] [PDF]

				K. Miwa, N. Yoshida, K. Nakagawa, and H. Inoue High-density lipoprotein particles are large in patients with variant angina Cardiovasc Res, March 1, 1998; 37(3): 729 - 737. [Abstract] [Full Text] [PDF]

This Article Abstract Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Permissions Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Syvänne, M. Articles by Taskinen, M.-R. Search for Related Content PubMed PubMed Citation Articles by Syvänne, M. Articles by Taskinen, M.-R.