(Circulation. 1997;95:825-830.)
© 1997 American Heart Association, Inc.
Articles |
the Cardiovascular Division (F.K.W., M.A.M., P.A.S., T.H.M., I.L., J.E.M., G.H.T.), Institute for the Prevention of Cardiovascular Disease, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Mass; and Framingham Heart Study (P.W.F.W., D.L.), Framingham, Mass.
Correspondence to Dr Francine K. Welty, One Autumn St, 5th Floor, Boston, MA 02215. E-mail fwelty@nedhmail.nedh.harvard.edu.
| Abstract |
|---|
|
|
|---|
Methods and Results Hemostatic risk factors were measured in 1878 individuals (1003 women and 875 men) participating in cycle 5 of the Framingham Offspring Study. The subjects were divided into five groups on the basis of LDL cholesterol level. Subjects with hypobetalipoproteinemia (LDL cholesterol <70 mg/dL) had the lowest levels of fibrinogen, plasminogen activator inhibitor1 antigen, and tissue plasminogen activator antigen. As LDL cholesterol increased, there was a significant increase in the levels of the hemostatic risk factors, with the exception of von Willebrand factor antigen. Adjustment with multivariate regression analyses for the covariates age, sex, body mass index, diabetes mellitus, smoking, alcohol intake, triglyceride level, and use of antihypertensive medication did not materially alter the results.
Conclusions Decreasing levels of LDL cholesterol are associated with decreasing levels of hemostatic risk factors. Subjects with hypobetalipoproteinemia have the lowest levels of hemostatic risk factors and may be protected against thrombotic complications of atherosclerotic cardiovascular disease because of reduced thrombotic potential. One mechanism by which lipid-lowering therapy may decrease clinical cardiac events is through a reduction in thrombotic tendency.
Key Words: cholesterol fibrinogen lipids plasminogen activator inhibitor von Willebrand factor
| Introduction |
|---|
|
|
|---|
Hypobetalipoproteinemia is characterized by plasma concentrations of apo B and LDL cholesterol that are one third of those observed in the general population.19 Although the relation between decreased cholesterol levels and thrombotic potential is amenable to study in subjects with hypobetalipoproteinemia, there have been, to our knowledge, no studies of hemostatic variables in these individuals. The aim of this study was to utilize subjects with hypobetalipoproteinemia to examine the relation between thrombotic potential and low levels of LDL cholesterol. We measured levels of fibrinogen, PAI-1 antigen, TPA antigen, vWF, and factor VII in subjects involved in the Framingham Offspring Study.
| Methods |
|---|
|
|
|---|
Individuals who reported smoking at least one cigarette per day during the year before the index examination were classified as current smokers. Body mass index was computed by dividing the weight (kilograms) by the square of the height (meters). Diabetes mellitus was considered present if fasting blood glucose exceeded 140 mg/dL or if insulin or oral hypoglycemic agents were used. Hypertension was defined as a blood pressure of >140/90 mm Hg or the use of antihypertensive medication. Alcohol consumption was assessed as the reported number of drinks per day of beer (12 oz), wine (4 oz), and spirits (1 oz).
Subjects with established coronary artery disease or stroke were excluded from the analysis (n=176); none of these had hypobetalipoproteinemia. A total of 1878 subjects (1003 women and 875 men) met entry criteria and were divided into five categories on the basis of LDL cholesterol for the present analysis. The first category included those with hypobetalipoproteinemia (LDL cholesterol level of <70 mg/dL, which is less than the fifth percentile for age and sex), and the second category was defined as an LDL cholesterol level of 70 to 100 mg/dL. The remaining categories were based on the National Cholesterol Education Program cutpoints: LDL cholesterol levels of 101 to 129, 130 to 159, and
160 mg/dL.
Blood Sampling and Hemostatic Analyses
Blood samples were collected from an antecubital vein between 8:00 and 9:00 AM after an overnight fast. Blood was anticoagulated with 3.8% trisodium citrate (9:1 vol/vol) and kept on crushed ice until centrifugation. Plasma was separated by centrifugation at 2500g for 30 minutes at 4°C. Plasma aliquots were quickly frozen and stored at -70°C for subsequent analysis. PAI-1 antigen levels were determined with a commercially available ELISA according to the description of Declerk et al20 (TintElize PAI-1, Biopool AB). Fibrinogen levels were determined according to the method of Clauss.21 Levels of TPA antigen were measured using ELISA (TintElize TPA, Biopool AB) according to the procedure described by Ranby et al.22 Factor VII antigen levels were measured with an ELISA with a commercially available kit (Diagnostica Stago). vWF antigen levels were measured with an ELISA as described previously.23 The intra-assay coefficient of variation was 9.6% for PAI-1 antigen, 2.6% for fibrinogen, 5.5% for TPA antigen, 8.8% for vWF, and 3.0% for factor VII.
Lipid Analysis
For determination of lipids, blood was anticoagulated with EDTA at a final concentration of 1 mg/mL. Plasma was separated by centrifugation at 2500g for 30 minutes at 4°C, and lipid measurements were made in fresh specimens. HDL cholesterol was measured after precipitation of LDL cholesterol and VLDL cholesterol with dextran-magnesium.24 Plasma levels of total cholesterol, HDL cholesterol, and triglycerides were measured according to automated enzymatic methods with an Abbott Diagnostics ABA-200 bichromatic analyzer and Abbott A-Gent enzymatic reagents.25 The level of LDL cholesterol was calculated with the Friedewald equation in all cases with triglyceride levels of <400 mg/dL.26 The laboratory participates in the Centers for Disease Control and Prevention lipid standardization program.
Statistical Analysis
To investigate the association between LDL cholesterol and hemostatic risk factors, we used multiple regression analyses. Subjects were divided into five categories based on their LDL cholesterol levels as described above. A series of models were fit with each of the hemostatic risk factors as the dependent variables. Mean values are presented for each of the hemostatic risk factors, first unadjusted and then after multivariable adjustment for age, sex, body mass index, diabetes mellitus, smoking, alcohol intake, and use of antihypertensive medication. Logarithmic transformation of levels of PAI-1 antigen, TPA antigen, and triglycerides was performed to attain a normal distribution.
Baseline clinical characteristics across the five categories of LDL cholesterol were compared using ANOVA and Pearson's
2 test for continuous and discrete variables, respectively. All data are presented as mean±SEM unless otherwise noted. Geometric mean values are presented for PAI-1 antigen and TPA antigen. Values for two-tailed P<.05 were considered statistically significant.
| Results |
|---|
|
|
|---|
|
|
Hemostatic Risk Factors
Table 3
shows the unadjusted levels of hemostatic risk factors in subjects with hypobetalipoproteinemia compared with those with higher levels of LDL cholesterol. There was a significant increase in the levels of hemostatic risk factors with increasing LDL cholesterol, with the exception of vWF antigen. Levels of fibrinogen, PAI-1 antigen, TPA antigen, and vWF antigen were the lowest in subjects with hypobetalipoproteinemia.
|
The levels of hemostatic risk factors after adjustment for age, sex, body mass index, diabetes mellitus, smoking, alcohol intake, and use of antihypertensive medications are shown in Table 4
. Adjustment for these covariates did not materially alter the results. Levels of fibrinogen, PAI-1 antigen, TPA antigen, and vWF antigen were the lowest in subjects with hypobetalipoproteinemia compared with subjects in the other LDL cholesterol categories.
|
| Discussion |
|---|
|
|
|---|
The physiological significance of low cholesterol levels is not completely understood. The low levels of apo Bcontaining lipoproteins and LDL cholesterol in subjects with hypobetalipoproteinemia would be predicted to prevent coronary artery disease by limiting the size of plaque formation. The data we present suggest that additional factors affecting the onset of acute cardiovascular syndromes, such as the clotting system, are also favorably altered by low LDL cholesterol levels.
Although atherosclerotic cardiovascular disease would be expected to be rare in subjects with hypobetalipoproteinemia, it can occur. An autopsy of a subject with hypobetalipoproteinemia (caused by a truncated apo B-55) and a mean total cholesterol level of 126 mg/dL throughout his lifetime revealed severe atherosclerosis of the coronary arteries, aorta, and iliac vessels; however, he did not have clinical symptoms of coronary disease during his lifetime.27 He was not diabetic and never smoked. A second subject with hypobetalipoproteinemia caused by a truncated apo B developed coronary heart disease. The interaction with other potential risk factors responsible for the development of atherosclerosis in these individuals is not apparent.
Impaired Fibrinolysis, Enhanced Thrombosis, and Coronary Heart Disease
The importance of determining the effect of low lipid levels on factors promoting coagulation and enhancing the fibrinolytic system is demonstrated by epidemiological studies that demonstrate that fibrinogen, TPA antigen, and vWF antigen are independent risk factors for cardiovascular disease in healthy subjects13 and recent studies demonstrating that impaired fibrinolytic capacity is associated with an increase in cardiovascular risk.14 15 16 17
In prospective studies, higher plasma levels of fibrinogen have been related to future coronary events in initially healthy subjects28 29 30 31 and in those within the first year of myocardial infarction.32 ECAT recently demonstrated in patients with angina pectoris that high levels of fibrinogen, TPA antigen, and vWF antigen were independent predictors of myocardial infarction or sudden cardiac death.13 In patients with high serum cholesterol levels in the ECAT study, the risk of coronary events rose with increasing levels of fibrinogen, but the risk remained low in the presence of low fibrinogen concentrations.13 This raises the question of whether the fibrinogen level is an indicator of the extent of inflammation in the progression of coronary artery disease. The prospective Northwick Park Heart Study indicated that factor VII coagulant activity (factor VIIc) is independently associated with the risk of future coronary heart disease in middle-aged men.12 28
In most studies of fibrinolytic capacity, levels of PAI-1 antigen, which is an inhibitor of TPA antigen and therefore an inhibitor of fibrinolysis, have been measured.33 34 35 36 37 Several prospective studies have shown that patients with high levels of PAI-1 antigen after myocardial infarction had an increased risk for reinfarction.12 37 38 Recent evidence links elevated levels of TPA antigen to an increased risk of future myocardial infarction in asymptomatic men39 and in patients with severe angina,40 to an increased risk of stroke,41 and to increased mortality in patients with known coronary artery disease.42 Most of the TPA antigen measured is complexed with PAI-1 antigen and is inactive. Therefore, high levels of TPA antigen in association with high levels of PAI-1 antigen may reflect impairment of the fibrinolytic system.43 44 45 46 47 48 High TPA antigen levels may also reflect a compensatory attempt to increase fibrinolysis in response to chronic hypercoagulability occurring in association with atherosclerosis. In the present study, the low TPA antigen levels may be secondary to low levels of PAI-1 antigen and LDL cholesterol. This combination probably reflects a favorable fibrinolytic potential.
Lipids, Thrombosis, and Fibrinolysis
Elevated lipid levels are known to be associated with a thrombotic tendency. Epstein et al49 demonstrated an impaired response of fibrinolytic activity to exercise in patients with type IV hyperlipidemia. Carvalho et al50 demonstrated that platelets from patients with type II hyperlipidemia undergo aggregation at 1/25 the concentration of epinephrine and one third the concentration of collagen and ADP required to induce aggregation of platelets from individuals with normal lipid levels. Simpson et al51 demonstrated that fibrinolytic activity was lower in patients with type IV hyperlipidemia than in subjects with normal lipid levels. Type IV patients also had higher fibrinogen and factor X levels,51 and patients with type II hyperlipoproteinemia had raised fibrinogen levels and increased platelet aggregation.52 Rats fed high cholesterol diets had significantly higher fibrinogen levels than those fed a normal chow diet.53
Effect of Lipid Lowering on Thrombotic and Fibrinolytic Potential
Recent clinical trials have documented that intensive lipid lowering produces significant reductions in ischemia-related clinical cardiac events with only minimal regression of angiographically demonstrated coronary artery stenoses.1 2 3 4 5 6 7 8 9 10 11 The mechanisms responsible for the marked reduction in acute ischemic events compared with the small amount of regression produced are unknown. We postulated that intensive lipid-lowering therapy may exert a major beneficial effect by reducing thrombotic tendency.18 Several reports have provided support for this. Lowering of triglyceride levels with clofibrate in individuals with severe hypertriglyceridemia resulted in a reduction in fibrinogen, as well as in factors VIIc and Xc.51 Also, treatment with clofibrate decreased platelet aggregability in patients with hyperbetalipoproteinemia.54 In a recent study, PAI-1 activity decreased in response to lovastatin therapy.55 Using an in vitro model, Lacoste et al56 showed that treatment of patients with stable coronary artery disease with pravastatin inhibited platelet thrombus formation at high and low shear rates. Gemfibrozil, a lipid-lowering fibrate, has been shown to decrease plasma levels of PAI-1 in subjects with elevated triglycerides57 and to attenuate increased synthesis of PAI-1 mediated by platelet-associated growth factors in cultured cells and animals.58 Niacin lowers plasma cholesterol by decreasing the synthesis of apo B and decreasing VLDL production from the liver59 60 ; treatment with niacin accelerated the rate of lysis of whole blood clots in vitro.61 62 Incubation of Hep G2 cells with niacin resulted in a 72% decrease in accumulation of PAI-1 protein due to decreased synthesis of PAI-1 and a decrease in the concentration of PAI-1 mRNA.63 These data strengthen the link between lipid metabolism and thrombotic potential in the blood.
Low Cholesterol and Cerebral Hemorrhage
Several studies have observed an association between low cholesterol levels and intracerebral hemorrhage.64 65 66 In 1971, researchers in Japan found the first link between low cholesterol and mortality when they observed a 2.53-fold increase in the risk for intracerebral hemorrhage in those with a total cholesterol level of <130 mg/dL and a 1.98-fold increase in those with total cholesterol of 130 to 159 mg/dL compared with those with cholesterol levels of
160 mg/dL. In the Multiple Risk Factor Intervention Trial, cholesterol levels of <160 mg/dL were associated with a twofold increase in risk of intracerebral hemorrhage.65 In the Honolulu Heart Program, the relative risk for intracerebral hemorrhage among individuals with total cholesterol of <160 mg/dL was 2.55 (95% confidence interval, 1.58 to 4.12) compared with those with total cholesterol
160 mg/dL after controlling for age, blood pressure, serum uric acid, cigarette smoking, and alcohol consumption.66 The reasons for this association are unknown. There is some evidence that low cholesterol adversely affects erythrocyte fragility and may increase the development of arterionecrosis.67 68 Our data, which suggest that low LDL cholesterol is associated with lower thrombotic potential, may partially explain the higher rates of hemorrhagic stroke in subjects with very low cholesterol. Perhaps when a cerebral vessel ruptures, a clot is less likely to seal the rupture in subjects with very low cholesterol, and therefore, a clinical cerebrovascular event is more likely to occur.
Conclusions
The results in the present study show a decreased thrombotic potential in subjects with hypobetalipoproteinemia compared with subjects with higher levels of LDL cholesterol. Our data suggest that hypobetalipoproteinemia may be associated with decreased endothelial elaboration of PAI-1 antigen and that LDL cholesterol levels of <70 mg/dL are highly beneficial for the coagulation system. This may be a major mechanism protecting subjects with hypobetalipoproteinemia from coronary heart disease. Our findings suggest that one mechanism by which lipid-lowering therapy may markedly decrease clinical cardiac events is through a reduction in thrombotic tendency. This should be examined in a randomized study of lipid-lowering therapy. In addition, adjustment for LDL cholesterol should be considered when fibrinogen, PAI-1, TPA antigen, and factor VII are evaluated as independent risk factors for coronary heart disease.
| Selected Abbreviations and Acronyms |
|---|
|
| Acknowledgments |
|---|
Received August 12, 1996; revision received September 26, 1996; accepted October 7, 1996.
| References |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
C. T. Park and S. D. Wright Fibrinogen is a component of a novel lipoprotein particle: Factor H-related protein (FHRP)-associated lipoprotein particle (FALP) Blood, January 1, 2000; 95(1): 198 - 204. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Circulation Home | Subscriptions | Archives | Feedback | Authors | Help | AHA Journals Home | Search Copyright © 1997 American Heart Association, Inc. All rights reserved. Unauthorized use prohibited. |