- Email: [email protected]
Lipoprotein phenotype and adhesion molecules correlate with diurnal triglyceride profiles in patients with coronary artery disease
20
Nutr Metab Cardiovasc Dis (2004) 14:20-25
ORIGINAL ARTICLE
Lipoprotein phenotype and adhesion molecules correlate with diurnal triglyceride profiles in patients with coronary artery disease A. Schmidt-Trucksfiss, M.W. Baumstark, C. Daub, S. Espenschied, D. Grathwohl, and A. Berg Freiburg University Hospital, Center for Internal Medicine, Department of Rehabilitative and Preventive Sports Medicine, Freiburg, Germany
Abstract Background and aim: To assess the relationships between different diurnal triglyceride (TG) profiles (p) and the atherogenicity of the lipoprotein phenotype and adhesion molecule concentrations in patients with coronary artery disease (CAD). Methods and Results: Repeated measurements of fasting TG and TGp were made in 29 CAD patients; fasting cholesterol levels (total-C, VLDL, LDL, HDL and small dense LDL) and soluble cell adhesion molecules (sCAM) (ICAM-1 and E-selectin) were measured once. Three different TGps were defined:fasting (137.0+_60.7 mg/dL) and all other TG levels <200 mg/dL (LL; n=7); a fasting TG level <200 mg/dL (147.0+_49.9 mg/dL) and maximum TG levels >200 mg/dL (LH; n=13); and both fasting (225.1+_76.2 mg/dL) and maximum TG levels" >200 mg/dL (HH; n=9). We then analysed the associations between the TGp types and the lipoprotein phenotype and CAMs'. LL had significantly lower values than LH (p
Key words: Coronary artery disease, E-selectin, hypertriglyceridemia, ICAM, small dense LDL. Correspondence to: Arno Schmidt-Trucksfiss, MD, Freiburg University Hospital, Center for Internal Medicine, Department of Rehabilitative and Preventive Sports Medicine, Hugstetter Str. 55, 79106 Freiburg, Germany E-mail: schmidt@msml'ukl'uni'freiburg'de Received: 10 October 2002; accepted: 9 September 2003
(209.4+_30.3, 267.5+_60.6, 273.4+_59.1 ng/dL) and sE-selectin (25.1+_17.6, 35.5+_11.5, 48.5+_20.2 ng/dL). Conclusion: Although the differences in fasting TG levels between the LL and LH groups were not significantly different, LH had a more atherogenic lipoprotein phenotype and higher concentrations of adhesion molecules. TGp measurements seem to be suitable for identifying CAD patients with an unfavourable diurnal TG and atherosclerosis-prone lipoprotein metabolism. Nutr Metab Cardiovasc Dis (2004) 14:20-25 ~2004, Medikal Press
Introduction Over the past 40 years, hypertriglyceridemia has been discussed as an independent risk factor for atherosclerosis and the subsequent manifestation of major cardiovascular diseases (1-3). High fasting triglyceride (TG) levels in conjunction with high small dense low-density lipoprotein (LDL) cholesterol and low high-density lipoprotein (HDL) cholesterol concentrations currently seem to be the best indicators of the risk of coronary artery disease (CAD). However, postprandial lipidemia may be an even better indicator of atherogenicity (4), because the longer the endothelium is exposed to TG-rich lipoproteins (remnants), the greater the deleterious effects on the vascular inner cell layer. Increased levels of cell adhesion molecules (CAMs) such as sE-selectin and sICAM-] have been associated with an unfavourable lipoprotein phenotype (5). CAMs
TGp types, lipoprotein phenotype and CAMs
play an important role in the development of atherosclerosis by initiating the adhesion of monocytes to endothelial cells, which leads to their subsequent transmigration into the intimal layer(6). Both the unfavourable lipoprotein phenotype (especially increased TG and small dense LDL levels, and reduced HDL concentrations) and the increased expression of cell adhesion molecules seem to act in the same direction and promote the atherosclerotic vascular process. The progression of this process should be prevented especially in patients at high risk of future cardiovascular events, such as those with prior myocardial infarction. In this respect, hypertriglyceridemia in the fasting and postprandial states seems to be a worthwhile target. Its measurement involves assessing fasting TG levels and usually a standardised oral fat test (4, 7). However, the oral fatloading test is not widely used, possibly because of the time required to administer it (up to eight hours) and the fact that patients often fail to avoid additional food intake during this period. This difficulty could be overcome by using a test procedure that is better suited to the subjects' everyday habits. In analogy with the diurnal glucose profile, we measured serum TG levels several times a day in non-hospitalised patients with manifest CAD in order to investigate whether their triglyceride profile (TGp) identifies CAD patients with prolonged postprandial TG clearance and providing additional information regarding the atherosclerotic risk of high TG levels. We therefore determined the corresponding lipoprotein phenotype and serum CAMs (slCAM-1 and sE-selectin) in hypertriglyceridemia. Our hypothesis was that the so-called lipid triad (high TG and small dense low-density lipoprotein levels, and low high-density lipoprotein levels) (3), and the concentrations of slCAM-1 and sE-selectin, are higher in patients with a less favourable TGp.
Methods Twenty-nine consecutive CAD outpatients (8 women and 21 men) participated in the study. Coronary angiography showed that eight had one-vessel, nine two-vessel, and twelve three-vessel disease. Fourteen had experienced a myocardial infarction and ten coronary bypass surgery at least three months before the study. Three presented peripheral arterial occlusive disease with a symptom-flee walking distance of more than 500 metres. All of the patients were clinically stable without any clinical signs of
21
cardiac insufficiency, and their symptom-flee maximum ergometric work capacity was >1.0 watt/kg body weight.The risk factor profile is shown in Table 2. Nineteen patients were regular alcohol consumers, with a daily alcohol intake of 23.9 g/d (LL 18.8 g/d: n=6; LH 25.4 g/d: n=7; HH 27.3 g/d: n=6). The study design was explained in detail to the participants, all of whom gave their written consent. The procedures used were in accordance with the Declaration of Helsinki of 1975, as revised in 1983.
Laboratory tests Blood samples were taken from the antecubital vein with the patients in a supine position after a fasting period of at least twelve hours. Serum glucose levels were analysed using a standard laboratory kit (Roche Diagnostics - Mannheim). Lipoproteins were isolated by means of sequential preparative ultracentrifugation at the following densities: less than 1.006 kg/L for VLDL; between 1.006 and 1.019 kg/L for intermediate density lipoprotein (IDL); between 1.019 and 1.063 kg/L for LDL; and between 1.063 and 1.21 for HDL. LDL was separated into six and HDL into three subclasses by means of equilibrium density-gradient ultracentrifugation as previously described (8). The density intervals for the LDL subclasses as determined by precision refractometry of blank gradients were: LDL-I: 1.019-1.031 kg/L; LDL-2: 1.031-1.034 kg/L; LDL-3:1.034-1.037 kg/L; LDL-4: 1.0371.040 kg/L; LDL-5:1.040-1.044 kg/L; and LDL-6: 1.0441.063 kg/L. The density intervals of the HDL subfractions were HDL2b: 1.063-1.100 kg/L; HDL2a: 1.100-1.125 kg/L; and HDL3:1.125-1.210 kg/L (9). Cholesterol levels in the LDL and HDL subclasses were analysed using an automated (EPOS; Eppendorf, Hamburg, Germany) enzymatic method (Boehringer Mannheim, Germany); apolipoprotein-B levels in the LDL subclasses were measured using endpoint nephelometry (Behring, Marburg, Germany), standardised according to the Center of Disease Control (CDC). In order to measure slCAM-1 and sE-selectin levels, the blood samples were centrifuged for 20 minutes at 4000 rpm, and then immediately frozen at -20°C. After all of the samples had been collected, the levels of slCAM-1 and sE-selectin were determined by means of an enzymatic immune assay (R&D Systems, Minneapolis, USA). Once the fasting blood samples had been taken, all of the participants underwent an oral glucose tolerance test (75 g defined mixture of mono- and oligosaccharides in 300 mL of water; DEXTRO ® O.G-T., Boehringer
A. Schmidt-Trucks&ss,et al
22
Mannheim). Blood glucose levels were measured photometrically (Accutrend GCT; Roche Diagnostics, Mannheim, Germany) before ingestion, and then after 30, 60, 90 and 120 minutes. On the basis of the WHO guidelines, impaired glucose tolerance was defined as a serum glucose level of more than 140-200 mg/dL (7.8-11.1 mmol/L) after 120 minutes. Values of >200 rag/alL (>11.1 retool/L) were defined as diabetes.
Assessment of diurnal triglyceride profiles After having been trained to do so by a medical professional for two days, the patients measured their own diurnal TG profile from finger tip blood on 4-5 occasions over a 6-month period using the Accutrend GCT. They were advised to wash their hands with warm water before carrying out the blood test and, adhering strictly to the instructions, shown how to obtain the drop of blood in such a way as to avoid excessive "milking" (10). When laypersons comply with these measures, the results are comparable with those of professional testers (10). The patients were asked to measure their TG levels six times per day: in the early morning (fasting value), and then at 9 and 12 a.m., and 3, 6 and 9 p.m. The patients recorded the time of the measurements, and there were only rare deviations of more than one hour due to individual or organisational reasons. The 12 a.m. measurement had to be made before lunch. The patients were asked to take notes concerning their food intake on the days the TGps were measured, and advised to follow their usual diet and adhere to the NCEP step II diet guidelines. On the basis of the American Heart Association's guidelines with an upper fasting TG limit of 200 mg/dL (borderline-high), three different TGps were defined: fasting and all other TG levels <200 mg/dL (LL), a fasting TG level <200 mg/dL and maximum TG level >200 mg/dL (LH), and both fasting and maximum TG levels >200 mg/dL (HH). The patients were assigned to a TGp type on the basis of the majority of their TGps.
Statistical analysbs The average mean TGps (±SD) were used for the descriptive analysis, and their differences tested by means of the non-parametric Mann-Whitney U Test. Holm's correction procedure was used to correct for multiple testing (11). The relationships between the parameters were revealed using Spearman's correlation coefficient (r). A p value of <0.05 was considered significant, and one of <0.01 highly significant. All of the data were calculated using the SPSS software program for Windows, version 7.5.2.
TABLE 1
Anthropometric data, systolic(SBP) and diastolic blood pressure (DBP), fasting and 120-minuteglucoselevel in the oral glucose tolerance test (OGTT) of all study subjectsin the subgroups according to the diurnal triglycerideprofiles. Total N=29
LL N=7
LH N=13
HH N=9
Age (years)
62+9
65+10
61 ±11
62±6
Sex (female/male)
7/22
1/6
3/10
1/8
172±9
175±9
172±10
170+7
Body mass (kg)
77.6±10.7
78.5±13.1
74.8±6.5
81.1±13.6
BMI (kg/m 2)
26.2±3.3
25.4+2.8
25.6±3.6
27.8±3.1
SBP (mmHg)
126±20
136±21"
125±20
121±20
DBP (mmHg)
85±8
85±6
86±7
82_+10
99±21
94+18
109±38
148±102
118±62
151_+85
Height (cm)
Glucose (mg/dL)
100_+27
OG-FI- 120 (mg/dL)
136±78
*p<0.05 LL v s HH. Differences in TGps by the Mann-WhitneyU test corrected for multiple testing according to Holm.
Results Anthropometric data There were no significant differences between the TGp groups in terms of age, sex, height, body mass, body mass index (BMI), fasting glucose or 2-hour OGTT values, but systolic blood pressure was significantly higher in the LL group (Table 1).
Risk factors and medication The distributions of the risk factors were similar in terms of the absolute number of subjects in each group (Table 2). The overall percentage of former smokers was high in all groups. Statin treatment was significantly more frequent in the HH group. None of the subjects in the LL group was taking an angiotensin-converting enzyme (ACE) inhibitor. There was no significant between-group difference in beta-blocker intake (Table 2).
Triglyceride profiles The TG values in the LL group were similar throughout the day (range: 127+62 mg/dl to 145+53 mg/dL). The LH group was characterised by a moderate increase mainly between 12 a.m. and 3 p.m.: the average pre-noon values
TGp types, lipoprotein phenotype and CAMs
23
TABLE 2
Risk factorprofile and medicationof the studysubjectsas a whole and the subgroupsof diurnaltriglycerideprofiles.
Hypertension, without medication
Total N=29
LL N=7
LH N=13
HH N =9
18
4
9
5
Diabetes
7
2
4
1
Overweight, BMI >25 kg/m 2
9
2
4
3
Former smoker
20
5
11
4
Current smoker
4
-
2
2
Medication i3-blocker
23
4
10
7
Statin
21
3**
9§
9
Oral antidiabetic
7
1
4
1
Insulin
4
2
1
1
ACE-inhibitor
7
-*
3
4
Salicylic acid
28
7
12
9
*p<0.05 LL vs HH; **p<0.01 LL vs HH; §p<0.05LH v s HH by the Mann-Whitney U test corrected for multipletesting accordingto Holm.
ranged from 146±69 to 171___73mg/dL, whereas the afternoon values were about 250 mg/dL (maximum 270___108 mg/dL). In the HH group, the values tended to decrease during the morning (from 375+141 mg/dL to 297+142 mg/dL), and peaked in the evening (391___140mg/dL). Most of the maximum values were measured after noon (69%), with the highest values being recorded at 3 p.m.; only 17% were recorded between 10 and 12 a.m., and only 14% were fasting values. The relationship between fasting and peak TG values was moderately significant (r=0.64, p<0.01), but this correlation is not strong enough to draw any firm conclusions. Stratification of the different TGps clearly showed that the fasting value was <200 mg/dL and the peak >200 mg/dL in nearly 50% of the cases.
Lipoprotein phenotype There were no significant differences among the TGp types in terms of total cholesterol, LDL and HDL (Table 3). Total cholesterol and LDL did not correlate with fasting TG levels or TGp types; HDL just failed to reach significance (r=0.32;p=0.055). As expected, VLDL was absolutely (but not significantly)
Total N=29
LL N=7
LH N=13
HH N=9
Triglyceride (mg/L)
168.8±70.8
137.0±60.7"
147.0±49.9
375.1 ±141.3
TotaI-C (mg/dL)
194.6±39.8
191.1 ±38.3
185.2±26.3
210.5±54.4
Total LDL-C (mg/dL)
90.8±27.7
90.7±28.2
86.4±19.4
97.1 ±38.0
LDL-C1 (mg/dL)
17.5±9.4
21.6±9.6
15.9±7.6
16.6±11.5
LDL-C2 (mg/dL)
10.6±7.0
14.9±8.6
9.9±6.4
8.4±5.6
LDL-C3 (mg/dL)
12.7±7.0
16.1 ±7.6
12.2±5.9
10.7±7.9
LDL-C4 (mg/dL)
14.4±6.3
14.5±5.2
14.9±5.2
13.5±8.7
LDL-C5 (mg/dL)
17.0±6.7
11.6±3.3#**
16.4±4.5
22.1 ±7.9
LDL-C6 (mg/dL)
18.5±8.4
12.0±3.2#**
17.0±5.7§
25.7±9.6
IDL-C (mg/dL)
8.7±5.6
6.9±3.7
7.1 ±3.4
12.0±7.7
VLDL-C (mg/dL)
19.7+15.9
11.2±5.8"~
18.8±9.4
28.1 ±8.8
HDL-C (mg/dL)
46.4±17.3
50.3±12.2
46.4±22.7
43.2±11.8
13.1 ±5.3
15.4±6.9
12.3±6.0
12.5±5.6
HDL-C2A (mg/dL) HDL-C2B (mg/dL)
10.9±5.2
13.8±4.9
9.8±5.8
10.3±4.3
HDL-C 3 (mg/dL)
19.4±4.4
20.1 ±3.5
18.2±4.3
20.6+5.1
slCAM-1 (ng/mL)
254.2±58.6
209.4±30.3*§
267.5±60.6
273.4±59.1
sE-selectin (ng/mL)
37.3±17.9
25.1 ±17.6*
35.5±11.5
48.5±20.2
#p<0.05 LL v s LH, *p<0.05 LL v s HH, **p<0.01 LL v s HH, §p<0.05 LH v s HH.
TABLE 3
Lipoproteinphenotypeand soluble cell adhesionmolecules(ICAM-1 and E-selectin)of the studysubjects as a wholeand the subgroupsof diurnal triglycerideprofiles.
A. Schmidt-Trucks~ss, et al
24
lower in the LL group than in the LH group, whereas the difference between both groups and the HH group was significant (p<0.0l). VLDL significantly correlated with fasting TG levels and TGp (both: r=0.63,p<0.01) The levels of the LDL subclasses, atherogenic LDL-5 and LDL-6, were significantly higher in the LH group than in the LL group (p<0.05), but the highest values were found in the HH group (p<0.05 v e r s u s LH and p<0.01 v e r s u s LL) (Table 3). LDL-6 correlated with fasting TG (r=0.57, p<0.01), and both LDL-5 and LDL-6 significantly correlated with the TGp types (r=0.62, p<0.0t and r= 0.64, p<0.0l, respectively). The cascade of VLDL to LDL-6 apoB showed an increased concentration of the LDL precursors VLDL- and IDL apoB and significantly increasing concentrations of LDL-5 and LDL-6 apoB as the TGp subgroups became more unfavourable (Fig. 1). Adhesion
molecules
The concentration of sICAM-1 was significantly lower in the LL than in the LH and HH groups, and the concentration of sE-selectin was significantly lower in the LL than in the LH group (Table 3). sE-selectin significantly correlated with fasting TG levels (r=0.38, p<0.05) and even more so with TGp (r=0.54,p<0.01). Interestingly, sE-selectin correlated positively with LDL-5 and LDL-6 (both: r= 0.66,p<0.05).
FIGURE 1
Apo-B lipoprotein concentrations along the delipidation cascade from VLDL to LDL-6. #p<0.05LL vs LH, #~p<0.01LL vs LH, *p<0.05 LL vs HH, **p<0.01LL vs HH ~p<0.05LH vs HH 18-
16#
§
," /
c
._o 12
"
~ 1 0 ## o c 0 0 m 0
8-
<
6-
N
LH
,'/
I\\*
- t
..
'
4-
Discussion Postprandial hypertriglyceridemia is thought to be an independent variable affecting various risk factors for major arterial occlusive diseases (1, 2). Measuring of the diurnal triglyceride profile in a way similar to that proposed by Castro et al (12) might be a useful alternative to oral fat test (7). Although it is less standardised, it is probably more suitable for outpatients. In relation to the TGp subgroups defined in this study, fasting TG levels were barely 10% higher in the LH than in the LL group, whereas the postprandial values were almost twice as high, It therefore seems that fasting levels alone are insufficient to judge hypertriglyceridemia as a risk factor, as they masked the unfavourable diurnal profile in nearly 50% of the study subjects (the LH group). The clinical relevance of the atherogenic potential of this finding in the LH group is reinforced by the results of the analysis of the associated lipoprotein variables in the TGps. The LH group had significantly higher plasma levels of the LDL-5 and LDL-6 fractions than the LL group, and apo-B concentrations indicate an increasing number of TG-rich particles and small dense LDL particles (LDL-5 and -6). This is in line with recent findings concerning the small dense LDL phenotype and fasting (13), as well as postprandial lipidemia in men (14). The CAM results showed that the CAD patients with only borderline-high fasting TG levels (LH) had a significantly increased concentration of soluble ICAM-1 (within the range of the patients with hypertriglyceridemia at the times of all of the diurnal measurements) and an unfavourable lipoprotein phenotype. These results are consistent with a previous in v i t r o study showing that high postprandial TG levels may enhance the expression of ICAM-1 in human umbilical vein endothelial cells (15). The sEselectin results were similar, although the levels were not significantly higher in the LH group in comparison with the LL group. The increased levels of slCAM and sE-selectin further support the atherogenic nature of an unfavourable TGp. However, as we did not measure C-reactive protein levels, our data are not sufficient to answer the question as to whether the increase in CAMs is caused directly by small dense LDL (16) or mediated by the activation of proinflammatory mechanisms, probably through the NFkappaB pathway (17). Limitations
2 I
I
I
VLDL
IDL
LDL-1
I
I
I
I
LDL-2 LDL-3 LDL-4 LDL-5 LDL-6
The lack of a detailed dietary analysis meant that we could not determine whether the differences in lipoprotein phenotypes and CAMs between the LH and LL groups
TGp types, lipoprotein phenotype and CAMs
were primarily due to impaired postprandial TG clearance or differences in fat intake. p-blockers without any intrinsic sympathomimetic activity can increase TG levels by up to 20% and reduce HDL by up to 5% (18), and may delay postprandial TG clearance (19). However, their effect on TGps may be less as their distribution was similar in the three groups. Alcohol consumption can increase serum TG levels. In one study, an acute intake of alcohol (38 ml ethanol in water) caused a transient increase in TG in normolipidemic subjects of 53% six hours later, whereas patients with known hypertriglyceridemia experienced an increase of only 3% (20). As alcohol consumption was comparable (especially between LL and LH groups), at least the range of its effects on the TGps should be similar, and thus not significantly influence the interpretation of the data. As this was an observational study of a small population, its results cannot be generalised to the population as a whole. In conclusion, measuring diurnal triglyceride profiles helps detecting unfavourable serum triglyceride levels during the day in patients with CAD more than measuring fasting triglyceride concentrations.The adverse nature of a TGp with fasting values of less than 200 mg/dL and a diurnal peak of more than 200 mg/dL is strongly supported by the associated unfavourable lipoprotein phenotype. The atherogenicity of this TGp type is further supported by the increased concentrations of sICAM-1 and sE-selectin. Although we do not propose using TGp measurements to replace the standardised tests for diagnosing impaired postprandial triglyceride clearance, they may be better suited to everyday clinical practice and thus be a practical alternative means of identifying patients with hypertriglyceridemia.
References 1. Hokanson JE, Austin MA (1996) Plasma triglyceride level is a risk factor for cardiovascular disease independent of high-density lipoprotein cholesterol level: a meta-analysis of population-based prospective studies. J Cardiovasc Risk 3:213-219 2. Gotto AM, Jr (1998) Triglyceride as a risk factor for coronary artery disease. Am J Cardio182: 22Q-25Q 3. Grundy SM (1998) Hypertriglyceridemia, atherogenic dyslipidemia, and the metabolic syndrome. Am J Cardiol 81: 18B-25B 4. Patsch JR, Miesenbock G, Hopferwieser T, Muhlberger V, Knapp E, Dunn JK, Gotto AM, Jr., Patsch W (1992) Relation of triglyceride metabolism and coronary artery disease. Studies in the postprandial state. Arterioscler Thromb 12:1336-1345 5. Hwang SJ, Ballantyne CM, Sharrett AR, Smith LC, Davis CE, Gotto AM, Jr, Boerwinkle E (1997) Circulating adhesion mole-
25
cules VCAM-1, ICAM-1, and E-selectin in carotid atherosclerosis and incident coronary heart disease cases: the Atherosclerosis Risk In Communities (ARIC) study. Circulation 96:4219-4225 6. Springer TA (1994) Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76:301-314 7. Groot PH, van Stiphout WA, Krauss XH, Jansen H, van Tol A, van Ramshorst E, Chin-On S, Hofman A, Cresswell SR, Havekes L (1991) Postprandial lipoprotein metabolism in normolipidemic men with and without coronary artery disease. Arterioscler Thromb 11:653-662 8. Baumstark MW, Kreutz W, Berg A, Frey I, Keul J (1990) Structure of human low-density lipoprotein subfractions, determined by X-ray small-angle scattering. Biochim Biophys Acta 1037:48-57 9. Anderson DW, Nichols AV, Forte TM, Lindgren FT (1977) Particle distribution of human serum high density lipoproteins. Biochim Biophys Acta 493:55-68 10. Luley C, Ronquist G, Reuter W, Paal V, Gottschling HD, Westphal S, King GL, Bakker SJ, Heine RJ, Hattemer A (2000) Point-of-care testing of triglycerides: evaluation of the Accutrend triglycerides system. Clin Chem 46:287-291 11. Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Statist 6:65-70 12. Castro CM, Halkes CJ, Meijssen S, van Oostrom A J, Erkelens DW (2001) Diurnal triglyceride profiles: a novel approach to study triglyceride changes. Atherosclerosis 155:219-228 13. Halle M, Berg A, Baumstark MW, Konig D, Huonker M, Keul J (1999) Influence of mild to moderately elevated triglycerides on low density lipoprotein subfraction concentration and composition in healthy men with low high density lipoprotein cholesterol levels. Atherosclerosis 143:185-192 14. Lemieux I, Couillard C, Pascot A, Bergeron N, Prud'homme D, Bergeron J, Tremblay A, Bouchard C, Mauriege P, Despres JP (2000) The small, dense LDL phenotype as a correlate of postprandial lipemia in men. Atherosclerosis 153:423-432 15. Moers A, Fenselau S, Schrezenmeir J (1997) Chylomicrons induce E-seleetin and VCAM-1 expression in endothelial cells. Exp Clin Endocrinol Diabetes 105: (Suppl 2) 35-37 16. Ambrosch A, Muller R, Freytag C, Borgmann S, Kraus J, Dierkes J, Neumann KH, Konig W (2002) Small-sized low-density lipoproteins of subclass B from patients with end-stage renal disease effectively augment tumor necrosis factor-alpha-induced adhesive properties in human endothelial cells. Am J Kidney Dis 39:972-984 17. Libby P (2002) Inflammation in atherosclerosis. Nature 420:868-874 18. Klein W (1992) Antihypertensive therapy and modification of metabolic risk factors (glucose and lipid metabolism). Z Kardiol 81:295-302 19. Barboriak JJ, Friedberg HD (1973) Propranolol and hypertriglyeeridemia. Atherosclerosis 17:31-35 20. Pownall HJ, Ballantyne CM, Kimball KT, Simpson SL, Yeshurun D, Gotto AM, Jr (1999) Effect of moderate alcohol consumption on hypertriglyceridemia: a study in the fasting state. Arch Intern Med 159:981-987
Nutr Metab Cardiovasc Dis (2004) 14:20-25
ORIGINAL ARTICLE
Lipoprotein phenotype and adhesion molecules correlate with diurnal triglyceride profiles in patients with coronary artery disease A. Schmidt-Trucksfiss, M.W. Baumstark, C. Daub, S. Espenschied, D. Grathwohl, and A. Berg Freiburg University Hospital, Center for Internal Medicine, Department of Rehabilitative and Preventive Sports Medicine, Freiburg, Germany
Abstract Background and aim: To assess the relationships between different diurnal triglyceride (TG) profiles (p) and the atherogenicity of the lipoprotein phenotype and adhesion molecule concentrations in patients with coronary artery disease (CAD). Methods and Results: Repeated measurements of fasting TG and TGp were made in 29 CAD patients; fasting cholesterol levels (total-C, VLDL, LDL, HDL and small dense LDL) and soluble cell adhesion molecules (sCAM) (ICAM-1 and E-selectin) were measured once. Three different TGps were defined:fasting (137.0+_60.7 mg/dL) and all other TG levels <200 mg/dL (LL; n=7); a fasting TG level <200 mg/dL (147.0+_49.9 mg/dL) and maximum TG levels >200 mg/dL (LH; n=13); and both fasting (225.1+_76.2 mg/dL) and maximum TG levels" >200 mg/dL (HH; n=9). We then analysed the associations between the TGp types and the lipoprotein phenotype and CAMs'. LL had significantly lower values than LH (p
Key words: Coronary artery disease, E-selectin, hypertriglyceridemia, ICAM, small dense LDL. Correspondence to: Arno Schmidt-Trucksfiss, MD, Freiburg University Hospital, Center for Internal Medicine, Department of Rehabilitative and Preventive Sports Medicine, Hugstetter Str. 55, 79106 Freiburg, Germany E-mail: schmidt@msml'ukl'uni'freiburg'de Received: 10 October 2002; accepted: 9 September 2003
(209.4+_30.3, 267.5+_60.6, 273.4+_59.1 ng/dL) and sE-selectin (25.1+_17.6, 35.5+_11.5, 48.5+_20.2 ng/dL). Conclusion: Although the differences in fasting TG levels between the LL and LH groups were not significantly different, LH had a more atherogenic lipoprotein phenotype and higher concentrations of adhesion molecules. TGp measurements seem to be suitable for identifying CAD patients with an unfavourable diurnal TG and atherosclerosis-prone lipoprotein metabolism. Nutr Metab Cardiovasc Dis (2004) 14:20-25 ~2004, Medikal Press
Introduction Over the past 40 years, hypertriglyceridemia has been discussed as an independent risk factor for atherosclerosis and the subsequent manifestation of major cardiovascular diseases (1-3). High fasting triglyceride (TG) levels in conjunction with high small dense low-density lipoprotein (LDL) cholesterol and low high-density lipoprotein (HDL) cholesterol concentrations currently seem to be the best indicators of the risk of coronary artery disease (CAD). However, postprandial lipidemia may be an even better indicator of atherogenicity (4), because the longer the endothelium is exposed to TG-rich lipoproteins (remnants), the greater the deleterious effects on the vascular inner cell layer. Increased levels of cell adhesion molecules (CAMs) such as sE-selectin and sICAM-] have been associated with an unfavourable lipoprotein phenotype (5). CAMs
TGp types, lipoprotein phenotype and CAMs
play an important role in the development of atherosclerosis by initiating the adhesion of monocytes to endothelial cells, which leads to their subsequent transmigration into the intimal layer(6). Both the unfavourable lipoprotein phenotype (especially increased TG and small dense LDL levels, and reduced HDL concentrations) and the increased expression of cell adhesion molecules seem to act in the same direction and promote the atherosclerotic vascular process. The progression of this process should be prevented especially in patients at high risk of future cardiovascular events, such as those with prior myocardial infarction. In this respect, hypertriglyceridemia in the fasting and postprandial states seems to be a worthwhile target. Its measurement involves assessing fasting TG levels and usually a standardised oral fat test (4, 7). However, the oral fatloading test is not widely used, possibly because of the time required to administer it (up to eight hours) and the fact that patients often fail to avoid additional food intake during this period. This difficulty could be overcome by using a test procedure that is better suited to the subjects' everyday habits. In analogy with the diurnal glucose profile, we measured serum TG levels several times a day in non-hospitalised patients with manifest CAD in order to investigate whether their triglyceride profile (TGp) identifies CAD patients with prolonged postprandial TG clearance and providing additional information regarding the atherosclerotic risk of high TG levels. We therefore determined the corresponding lipoprotein phenotype and serum CAMs (slCAM-1 and sE-selectin) in hypertriglyceridemia. Our hypothesis was that the so-called lipid triad (high TG and small dense low-density lipoprotein levels, and low high-density lipoprotein levels) (3), and the concentrations of slCAM-1 and sE-selectin, are higher in patients with a less favourable TGp.
Methods Twenty-nine consecutive CAD outpatients (8 women and 21 men) participated in the study. Coronary angiography showed that eight had one-vessel, nine two-vessel, and twelve three-vessel disease. Fourteen had experienced a myocardial infarction and ten coronary bypass surgery at least three months before the study. Three presented peripheral arterial occlusive disease with a symptom-flee walking distance of more than 500 metres. All of the patients were clinically stable without any clinical signs of
21
cardiac insufficiency, and their symptom-flee maximum ergometric work capacity was >1.0 watt/kg body weight.The risk factor profile is shown in Table 2. Nineteen patients were regular alcohol consumers, with a daily alcohol intake of 23.9 g/d (LL 18.8 g/d: n=6; LH 25.4 g/d: n=7; HH 27.3 g/d: n=6). The study design was explained in detail to the participants, all of whom gave their written consent. The procedures used were in accordance with the Declaration of Helsinki of 1975, as revised in 1983.
Laboratory tests Blood samples were taken from the antecubital vein with the patients in a supine position after a fasting period of at least twelve hours. Serum glucose levels were analysed using a standard laboratory kit (Roche Diagnostics - Mannheim). Lipoproteins were isolated by means of sequential preparative ultracentrifugation at the following densities: less than 1.006 kg/L for VLDL; between 1.006 and 1.019 kg/L for intermediate density lipoprotein (IDL); between 1.019 and 1.063 kg/L for LDL; and between 1.063 and 1.21 for HDL. LDL was separated into six and HDL into three subclasses by means of equilibrium density-gradient ultracentrifugation as previously described (8). The density intervals for the LDL subclasses as determined by precision refractometry of blank gradients were: LDL-I: 1.019-1.031 kg/L; LDL-2: 1.031-1.034 kg/L; LDL-3:1.034-1.037 kg/L; LDL-4: 1.0371.040 kg/L; LDL-5:1.040-1.044 kg/L; and LDL-6: 1.0441.063 kg/L. The density intervals of the HDL subfractions were HDL2b: 1.063-1.100 kg/L; HDL2a: 1.100-1.125 kg/L; and HDL3:1.125-1.210 kg/L (9). Cholesterol levels in the LDL and HDL subclasses were analysed using an automated (EPOS; Eppendorf, Hamburg, Germany) enzymatic method (Boehringer Mannheim, Germany); apolipoprotein-B levels in the LDL subclasses were measured using endpoint nephelometry (Behring, Marburg, Germany), standardised according to the Center of Disease Control (CDC). In order to measure slCAM-1 and sE-selectin levels, the blood samples were centrifuged for 20 minutes at 4000 rpm, and then immediately frozen at -20°C. After all of the samples had been collected, the levels of slCAM-1 and sE-selectin were determined by means of an enzymatic immune assay (R&D Systems, Minneapolis, USA). Once the fasting blood samples had been taken, all of the participants underwent an oral glucose tolerance test (75 g defined mixture of mono- and oligosaccharides in 300 mL of water; DEXTRO ® O.G-T., Boehringer
A. Schmidt-Trucks&ss,et al
22
Mannheim). Blood glucose levels were measured photometrically (Accutrend GCT; Roche Diagnostics, Mannheim, Germany) before ingestion, and then after 30, 60, 90 and 120 minutes. On the basis of the WHO guidelines, impaired glucose tolerance was defined as a serum glucose level of more than 140-200 mg/dL (7.8-11.1 mmol/L) after 120 minutes. Values of >200 rag/alL (>11.1 retool/L) were defined as diabetes.
Assessment of diurnal triglyceride profiles After having been trained to do so by a medical professional for two days, the patients measured their own diurnal TG profile from finger tip blood on 4-5 occasions over a 6-month period using the Accutrend GCT. They were advised to wash their hands with warm water before carrying out the blood test and, adhering strictly to the instructions, shown how to obtain the drop of blood in such a way as to avoid excessive "milking" (10). When laypersons comply with these measures, the results are comparable with those of professional testers (10). The patients were asked to measure their TG levels six times per day: in the early morning (fasting value), and then at 9 and 12 a.m., and 3, 6 and 9 p.m. The patients recorded the time of the measurements, and there were only rare deviations of more than one hour due to individual or organisational reasons. The 12 a.m. measurement had to be made before lunch. The patients were asked to take notes concerning their food intake on the days the TGps were measured, and advised to follow their usual diet and adhere to the NCEP step II diet guidelines. On the basis of the American Heart Association's guidelines with an upper fasting TG limit of 200 mg/dL (borderline-high), three different TGps were defined: fasting and all other TG levels <200 mg/dL (LL), a fasting TG level <200 mg/dL and maximum TG level >200 mg/dL (LH), and both fasting and maximum TG levels >200 mg/dL (HH). The patients were assigned to a TGp type on the basis of the majority of their TGps.
Statistical analysbs The average mean TGps (±SD) were used for the descriptive analysis, and their differences tested by means of the non-parametric Mann-Whitney U Test. Holm's correction procedure was used to correct for multiple testing (11). The relationships between the parameters were revealed using Spearman's correlation coefficient (r). A p value of <0.05 was considered significant, and one of <0.01 highly significant. All of the data were calculated using the SPSS software program for Windows, version 7.5.2.
TABLE 1
Anthropometric data, systolic(SBP) and diastolic blood pressure (DBP), fasting and 120-minuteglucoselevel in the oral glucose tolerance test (OGTT) of all study subjectsin the subgroups according to the diurnal triglycerideprofiles. Total N=29
LL N=7
LH N=13
HH N=9
Age (years)
62+9
65+10
61 ±11
62±6
Sex (female/male)
7/22
1/6
3/10
1/8
172±9
175±9
172±10
170+7
Body mass (kg)
77.6±10.7
78.5±13.1
74.8±6.5
81.1±13.6
BMI (kg/m 2)
26.2±3.3
25.4+2.8
25.6±3.6
27.8±3.1
SBP (mmHg)
126±20
136±21"
125±20
121±20
DBP (mmHg)
85±8
85±6
86±7
82_+10
99±21
94+18
109±38
148±102
118±62
151_+85
Height (cm)
Glucose (mg/dL)
100_+27
OG-FI- 120 (mg/dL)
136±78
*p<0.05 LL v s HH. Differences in TGps by the Mann-WhitneyU test corrected for multiple testing according to Holm.
Results Anthropometric data There were no significant differences between the TGp groups in terms of age, sex, height, body mass, body mass index (BMI), fasting glucose or 2-hour OGTT values, but systolic blood pressure was significantly higher in the LL group (Table 1).
Risk factors and medication The distributions of the risk factors were similar in terms of the absolute number of subjects in each group (Table 2). The overall percentage of former smokers was high in all groups. Statin treatment was significantly more frequent in the HH group. None of the subjects in the LL group was taking an angiotensin-converting enzyme (ACE) inhibitor. There was no significant between-group difference in beta-blocker intake (Table 2).
Triglyceride profiles The TG values in the LL group were similar throughout the day (range: 127+62 mg/dl to 145+53 mg/dL). The LH group was characterised by a moderate increase mainly between 12 a.m. and 3 p.m.: the average pre-noon values
TGp types, lipoprotein phenotype and CAMs
23
TABLE 2
Risk factorprofile and medicationof the studysubjectsas a whole and the subgroupsof diurnaltriglycerideprofiles.
Hypertension, without medication
Total N=29
LL N=7
LH N=13
HH N =9
18
4
9
5
Diabetes
7
2
4
1
Overweight, BMI >25 kg/m 2
9
2
4
3
Former smoker
20
5
11
4
Current smoker
4
-
2
2
Medication i3-blocker
23
4
10
7
Statin
21
3**
9§
9
Oral antidiabetic
7
1
4
1
Insulin
4
2
1
1
ACE-inhibitor
7
-*
3
4
Salicylic acid
28
7
12
9
*p<0.05 LL vs HH; **p<0.01 LL vs HH; §p<0.05LH v s HH by the Mann-Whitney U test corrected for multipletesting accordingto Holm.
ranged from 146±69 to 171___73mg/dL, whereas the afternoon values were about 250 mg/dL (maximum 270___108 mg/dL). In the HH group, the values tended to decrease during the morning (from 375+141 mg/dL to 297+142 mg/dL), and peaked in the evening (391___140mg/dL). Most of the maximum values were measured after noon (69%), with the highest values being recorded at 3 p.m.; only 17% were recorded between 10 and 12 a.m., and only 14% were fasting values. The relationship between fasting and peak TG values was moderately significant (r=0.64, p<0.01), but this correlation is not strong enough to draw any firm conclusions. Stratification of the different TGps clearly showed that the fasting value was <200 mg/dL and the peak >200 mg/dL in nearly 50% of the cases.
Lipoprotein phenotype There were no significant differences among the TGp types in terms of total cholesterol, LDL and HDL (Table 3). Total cholesterol and LDL did not correlate with fasting TG levels or TGp types; HDL just failed to reach significance (r=0.32;p=0.055). As expected, VLDL was absolutely (but not significantly)
Total N=29
LL N=7
LH N=13
HH N=9
Triglyceride (mg/L)
168.8±70.8
137.0±60.7"
147.0±49.9
375.1 ±141.3
TotaI-C (mg/dL)
194.6±39.8
191.1 ±38.3
185.2±26.3
210.5±54.4
Total LDL-C (mg/dL)
90.8±27.7
90.7±28.2
86.4±19.4
97.1 ±38.0
LDL-C1 (mg/dL)
17.5±9.4
21.6±9.6
15.9±7.6
16.6±11.5
LDL-C2 (mg/dL)
10.6±7.0
14.9±8.6
9.9±6.4
8.4±5.6
LDL-C3 (mg/dL)
12.7±7.0
16.1 ±7.6
12.2±5.9
10.7±7.9
LDL-C4 (mg/dL)
14.4±6.3
14.5±5.2
14.9±5.2
13.5±8.7
LDL-C5 (mg/dL)
17.0±6.7
11.6±3.3#**
16.4±4.5
22.1 ±7.9
LDL-C6 (mg/dL)
18.5±8.4
12.0±3.2#**
17.0±5.7§
25.7±9.6
IDL-C (mg/dL)
8.7±5.6
6.9±3.7
7.1 ±3.4
12.0±7.7
VLDL-C (mg/dL)
19.7+15.9
11.2±5.8"~
18.8±9.4
28.1 ±8.8
HDL-C (mg/dL)
46.4±17.3
50.3±12.2
46.4±22.7
43.2±11.8
13.1 ±5.3
15.4±6.9
12.3±6.0
12.5±5.6
HDL-C2A (mg/dL) HDL-C2B (mg/dL)
10.9±5.2
13.8±4.9
9.8±5.8
10.3±4.3
HDL-C 3 (mg/dL)
19.4±4.4
20.1 ±3.5
18.2±4.3
20.6+5.1
slCAM-1 (ng/mL)
254.2±58.6
209.4±30.3*§
267.5±60.6
273.4±59.1
sE-selectin (ng/mL)
37.3±17.9
25.1 ±17.6*
35.5±11.5
48.5±20.2
#p<0.05 LL v s LH, *p<0.05 LL v s HH, **p<0.01 LL v s HH, §p<0.05 LH v s HH.
TABLE 3
Lipoproteinphenotypeand soluble cell adhesionmolecules(ICAM-1 and E-selectin)of the studysubjects as a wholeand the subgroupsof diurnal triglycerideprofiles.
A. Schmidt-Trucks~ss, et al
24
lower in the LL group than in the LH group, whereas the difference between both groups and the HH group was significant (p<0.0l). VLDL significantly correlated with fasting TG levels and TGp (both: r=0.63,p<0.01) The levels of the LDL subclasses, atherogenic LDL-5 and LDL-6, were significantly higher in the LH group than in the LL group (p<0.05), but the highest values were found in the HH group (p<0.05 v e r s u s LH and p<0.01 v e r s u s LL) (Table 3). LDL-6 correlated with fasting TG (r=0.57, p<0.01), and both LDL-5 and LDL-6 significantly correlated with the TGp types (r=0.62, p<0.0t and r= 0.64, p<0.0l, respectively). The cascade of VLDL to LDL-6 apoB showed an increased concentration of the LDL precursors VLDL- and IDL apoB and significantly increasing concentrations of LDL-5 and LDL-6 apoB as the TGp subgroups became more unfavourable (Fig. 1). Adhesion
molecules
The concentration of sICAM-1 was significantly lower in the LL than in the LH and HH groups, and the concentration of sE-selectin was significantly lower in the LL than in the LH group (Table 3). sE-selectin significantly correlated with fasting TG levels (r=0.38, p<0.05) and even more so with TGp (r=0.54,p<0.01). Interestingly, sE-selectin correlated positively with LDL-5 and LDL-6 (both: r= 0.66,p<0.05).
FIGURE 1
Apo-B lipoprotein concentrations along the delipidation cascade from VLDL to LDL-6. #p<0.05LL vs LH, #~p<0.01LL vs LH, *p<0.05 LL vs HH, **p<0.01LL vs HH ~p<0.05LH vs HH 18-
16#
§
," /
c
._o 12
"
~ 1 0 ## o c 0 0 m 0
8-
<
6-
N
LH
,'/
I\\*
- t
..
'
4-
Discussion Postprandial hypertriglyceridemia is thought to be an independent variable affecting various risk factors for major arterial occlusive diseases (1, 2). Measuring of the diurnal triglyceride profile in a way similar to that proposed by Castro et al (12) might be a useful alternative to oral fat test (7). Although it is less standardised, it is probably more suitable for outpatients. In relation to the TGp subgroups defined in this study, fasting TG levels were barely 10% higher in the LH than in the LL group, whereas the postprandial values were almost twice as high, It therefore seems that fasting levels alone are insufficient to judge hypertriglyceridemia as a risk factor, as they masked the unfavourable diurnal profile in nearly 50% of the study subjects (the LH group). The clinical relevance of the atherogenic potential of this finding in the LH group is reinforced by the results of the analysis of the associated lipoprotein variables in the TGps. The LH group had significantly higher plasma levels of the LDL-5 and LDL-6 fractions than the LL group, and apo-B concentrations indicate an increasing number of TG-rich particles and small dense LDL particles (LDL-5 and -6). This is in line with recent findings concerning the small dense LDL phenotype and fasting (13), as well as postprandial lipidemia in men (14). The CAM results showed that the CAD patients with only borderline-high fasting TG levels (LH) had a significantly increased concentration of soluble ICAM-1 (within the range of the patients with hypertriglyceridemia at the times of all of the diurnal measurements) and an unfavourable lipoprotein phenotype. These results are consistent with a previous in v i t r o study showing that high postprandial TG levels may enhance the expression of ICAM-1 in human umbilical vein endothelial cells (15). The sEselectin results were similar, although the levels were not significantly higher in the LH group in comparison with the LL group. The increased levels of slCAM and sE-selectin further support the atherogenic nature of an unfavourable TGp. However, as we did not measure C-reactive protein levels, our data are not sufficient to answer the question as to whether the increase in CAMs is caused directly by small dense LDL (16) or mediated by the activation of proinflammatory mechanisms, probably through the NFkappaB pathway (17). Limitations
2 I
I
I
VLDL
IDL
LDL-1
I
I
I
I
LDL-2 LDL-3 LDL-4 LDL-5 LDL-6
The lack of a detailed dietary analysis meant that we could not determine whether the differences in lipoprotein phenotypes and CAMs between the LH and LL groups
TGp types, lipoprotein phenotype and CAMs
were primarily due to impaired postprandial TG clearance or differences in fat intake. p-blockers without any intrinsic sympathomimetic activity can increase TG levels by up to 20% and reduce HDL by up to 5% (18), and may delay postprandial TG clearance (19). However, their effect on TGps may be less as their distribution was similar in the three groups. Alcohol consumption can increase serum TG levels. In one study, an acute intake of alcohol (38 ml ethanol in water) caused a transient increase in TG in normolipidemic subjects of 53% six hours later, whereas patients with known hypertriglyceridemia experienced an increase of only 3% (20). As alcohol consumption was comparable (especially between LL and LH groups), at least the range of its effects on the TGps should be similar, and thus not significantly influence the interpretation of the data. As this was an observational study of a small population, its results cannot be generalised to the population as a whole. In conclusion, measuring diurnal triglyceride profiles helps detecting unfavourable serum triglyceride levels during the day in patients with CAD more than measuring fasting triglyceride concentrations.The adverse nature of a TGp with fasting values of less than 200 mg/dL and a diurnal peak of more than 200 mg/dL is strongly supported by the associated unfavourable lipoprotein phenotype. The atherogenicity of this TGp type is further supported by the increased concentrations of sICAM-1 and sE-selectin. Although we do not propose using TGp measurements to replace the standardised tests for diagnosing impaired postprandial triglyceride clearance, they may be better suited to everyday clinical practice and thus be a practical alternative means of identifying patients with hypertriglyceridemia.
References 1. Hokanson JE, Austin MA (1996) Plasma triglyceride level is a risk factor for cardiovascular disease independent of high-density lipoprotein cholesterol level: a meta-analysis of population-based prospective studies. J Cardiovasc Risk 3:213-219 2. Gotto AM, Jr (1998) Triglyceride as a risk factor for coronary artery disease. Am J Cardio182: 22Q-25Q 3. Grundy SM (1998) Hypertriglyceridemia, atherogenic dyslipidemia, and the metabolic syndrome. Am J Cardiol 81: 18B-25B 4. Patsch JR, Miesenbock G, Hopferwieser T, Muhlberger V, Knapp E, Dunn JK, Gotto AM, Jr., Patsch W (1992) Relation of triglyceride metabolism and coronary artery disease. Studies in the postprandial state. Arterioscler Thromb 12:1336-1345 5. Hwang SJ, Ballantyne CM, Sharrett AR, Smith LC, Davis CE, Gotto AM, Jr, Boerwinkle E (1997) Circulating adhesion mole-
25
cules VCAM-1, ICAM-1, and E-selectin in carotid atherosclerosis and incident coronary heart disease cases: the Atherosclerosis Risk In Communities (ARIC) study. Circulation 96:4219-4225 6. Springer TA (1994) Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm. Cell 76:301-314 7. Groot PH, van Stiphout WA, Krauss XH, Jansen H, van Tol A, van Ramshorst E, Chin-On S, Hofman A, Cresswell SR, Havekes L (1991) Postprandial lipoprotein metabolism in normolipidemic men with and without coronary artery disease. Arterioscler Thromb 11:653-662 8. Baumstark MW, Kreutz W, Berg A, Frey I, Keul J (1990) Structure of human low-density lipoprotein subfractions, determined by X-ray small-angle scattering. Biochim Biophys Acta 1037:48-57 9. Anderson DW, Nichols AV, Forte TM, Lindgren FT (1977) Particle distribution of human serum high density lipoproteins. Biochim Biophys Acta 493:55-68 10. Luley C, Ronquist G, Reuter W, Paal V, Gottschling HD, Westphal S, King GL, Bakker SJ, Heine RJ, Hattemer A (2000) Point-of-care testing of triglycerides: evaluation of the Accutrend triglycerides system. Clin Chem 46:287-291 11. Holm S (1979) A simple sequentially rejective multiple test procedure. Scand J Statist 6:65-70 12. Castro CM, Halkes CJ, Meijssen S, van Oostrom A J, Erkelens DW (2001) Diurnal triglyceride profiles: a novel approach to study triglyceride changes. Atherosclerosis 155:219-228 13. Halle M, Berg A, Baumstark MW, Konig D, Huonker M, Keul J (1999) Influence of mild to moderately elevated triglycerides on low density lipoprotein subfraction concentration and composition in healthy men with low high density lipoprotein cholesterol levels. Atherosclerosis 143:185-192 14. Lemieux I, Couillard C, Pascot A, Bergeron N, Prud'homme D, Bergeron J, Tremblay A, Bouchard C, Mauriege P, Despres JP (2000) The small, dense LDL phenotype as a correlate of postprandial lipemia in men. Atherosclerosis 153:423-432 15. Moers A, Fenselau S, Schrezenmeir J (1997) Chylomicrons induce E-seleetin and VCAM-1 expression in endothelial cells. Exp Clin Endocrinol Diabetes 105: (Suppl 2) 35-37 16. Ambrosch A, Muller R, Freytag C, Borgmann S, Kraus J, Dierkes J, Neumann KH, Konig W (2002) Small-sized low-density lipoproteins of subclass B from patients with end-stage renal disease effectively augment tumor necrosis factor-alpha-induced adhesive properties in human endothelial cells. Am J Kidney Dis 39:972-984 17. Libby P (2002) Inflammation in atherosclerosis. Nature 420:868-874 18. Klein W (1992) Antihypertensive therapy and modification of metabolic risk factors (glucose and lipid metabolism). Z Kardiol 81:295-302 19. Barboriak JJ, Friedberg HD (1973) Propranolol and hypertriglyeeridemia. Atherosclerosis 17:31-35 20. Pownall HJ, Ballantyne CM, Kimball KT, Simpson SL, Yeshurun D, Gotto AM, Jr (1999) Effect of moderate alcohol consumption on hypertriglyceridemia: a study in the fasting state. Arch Intern Med 159:981-987