Association of hemostatic factors with peripheral vascular disease

Peripheral Vascular Disease

Association of hemostatic factors with peripheral vascular disease Claire S. Philipp, MD, Laura A. Cisar, PhD, MPH, Hugh C. Kim, MD, Alan C. Wilson, PhD, Parvin Said|, MD, and J o h n B. Kostis, MD N e w Brunswick, NJ.

Hemostatic risk factors have been well established in coronary artery disease but less well studied in peripheral vascular disease. The relationship of coagulation and fibrinolytic proteins to lower limb arterial occlusive disease and other vascular risk factors remains poorly defined. Fibrinogen, factor VII coagulant activity, van Willebrand factor (vWf) antigen, and plasminogen activator inhibitor-1 (PAl-1) activity were measured in 46 adult participants in the Arterial Disease Multiple Intervention Trial (ADMIT) and in 76 control subjects and related to ankle-brachial systolic pressure index (ABI), a measure of lower limb arterial stenosis. The primary inclusion criterion for the ADMIT study population was on average of two ABIs <0.85. Fibrinogen and PAl-1 in ADMIT subjects were significantly higher than in control subjects (331 _+52 mg/dl vs 273 _ 46 mg/dl, p < 0.0001; 18.7 + 10 ~nits/ml vs 13.5 _+ 8.9 units/ml, p < 0.04). There were significant correlations of fibrinogen with ABI, factor VII coagulant activi~, and systolic and diastolic blood pressures; PAl-1 with body mass index and age; and factor VII coagulant activity with cholesterol levels. Logistic regression analysis, considering hemostatic variables and several known nonhemostatic risk fao 9 tars of peripheral arterial disease, showed that fibrinogen and systolic blood pressure were independently associated with ABI status in this population. The results demonstrate a strong independent correlation between fibrinogen levels and the presence of lower limb arterial stenosis. PAl-1 levels were elevated in ADMIT participants, but multivariate analysis did not demonstrate an independent relationship between PAl-1 and ABI. {Am Heart J 1997; 134:978-84.)

To date, abnormalities of the procoagulant and fibrinolytic systems have been described in association with coronary artery disease, 1-5 but hemostatic risk factors associated with lower limb arterial occlusive disease are not well established because only limited data have been reported in this population. 6 Epidemiologic evidence from a number of studies has linked several thrombogenic stimuli, including fibrinogen, plasminogen activator inhibitor-1 (PAl-D, factor VII, and von Willebrand factor (vWf), with ischemic heart disease and myocardial infarction. L-5,740 Experimental data support the role of these hemostatic proteins in both the process of thrombus formation and the development of proliferative atheromatous lesions. In addition to increasing blood viscosity and From the Divisionsof Hematologyand Cardiology, Universilyof Medicine and Dentistry of New Jersey-RobertWood JohnsonMedical School. Supported in part by the National Heart, Lungand fibod Institute Igrant no. NO 1HC-251151 and a grant from the Robert Wood JohnsonCardiovascularInstitute. SubrnlttedJan. 22, 1997;accepted June 20, 199Z Reprint requests: Claire S. Philipp, MD, Division of Hematobgy, UMDNJ-Robert Wood JohnsonMedical School, Medical Education Brdg., Room378, 1 Robert Wood JohnsonPL, New Brunswick, NJ 08904. Ecnail: [email protected] Copyright 9 1997 by Mosby-Year Book, Inc. 0002-8703/97/$5.00 + 0 411184856

participating in platelet aggregation and local thrombus formation, high concentrations of fibrinogen have been detected on the surface and within atherosclerotic plaques, n,la Elevated levels of PAI-1 have also been detected in the intima and media of atherosclerotic segments, 13 and it has been shown that increased plasma PAI-1 activity decreases fibrinolysis and increases thrombus extension) 4,15 Interaction of factor VIIa with tissue factor, which can be expressed on stimulated endothelial cells and some tissue macrophages, leads to tile generation of fibrin on exposed subendothelial surfaces. 16,17 In addition, decreased adhesion of platelets to subendothelium has been demonstrated with nonanticoagulated whole blood derived from homozygous vWf-deficient pigs, TMand reduced aortic atheromatous lesions have been found in these animals. 19 Clinical symptoms of peripheral arterial disease result from progressive lower limb arterial narrowing. 2~ In contrast to disease in the coronary circulation, lower extremity atherosclerotic narrowing often progresses gradually over long periods without causing critical symptoms, and it can be assessed noninvasively. An ankle-brachial systolic pressure index (ABI) of <0.9 correlates with disease on duplex scanning

Philipp et al.

a n d is >95% specific in d e t e c t i n g a n g i o g m m - p o s i t i v e p e r i p h e r a l arterial disease, w i t h p r o g r e s s i v e l y l o w e r ABIs indicative o f m o r e s e v e r e disease. 2~ O u r objective in this study was to d e t e r m i n e the relationship o f h e m o s t a t i c variables to atherosclerotic n a r r o w i n g a n d o t h e r v a s c u l a r risk factors w i t h the ABI u s e d as an indicator o f l o w e r limb arterial stenosis.

Methods Study population Aduhs, 30 years old or older, ``-ere uniformly contacted through random mailings and screened by the Cardiovascular Diseases and Hypertension Clinical Trials Center at our institution for participation as either a subject in the Arterial Disease Multiple Intervention Trial (ADMIT) or a control subject. Medical history, including past cardiovascular events and smoking history, was obtained on entry. Inclusion criteria for participation:as an ADMIT subject included an ABI <0.85, averaged frohl two screening visits, or documented previous surgery, angioplasty, or amputation for peripheral arterial disease. Control subjects ``ere required to be normotensive with systolic and diastolic blood pressures <140/90 mm Hg and to have an ABI >1.0 to i-educe the possibility of undetected peripheral vascular disease. Potential control subjects were excluded ff tliere was a history of peripheral vascular, cardiovascular, or cerebrovascular disease. Potential ADMIT subjects were excluded if there ``'as a history of myocardial infarction, stroke, or vascular surgery within the past 6 months; angioplasty in any arterial bed within the past 6 weeks; or unstable angina, congestive heart failure (New York Heart Association class III or IV), atrial fibrillation, or uncontrolled blood pressure (diastolic pressure >100 mm Hg or systolic pressure >180 mm Hg). Potential ADMIT and control subjects were excluded for low-density lipoprotein cholesterol levels >190 mg/dl, triglyceride levels >400 mg/dl, active cancer within the past 5 years, or poorly controlled diabetes (hemoglobin A l e >8.5% or a history of diabetic coma or ketoacidosis). Informed consent, which was approved by the University of Medicine and Dentistry of New Jersey Institutional Review Board, was obtained from all study participants.

Determination of ABI Left and right brachial systolic and diastolic blood pressures were taken in the supine position in all participants after 10 minutes of rest with a random-zero sphygmomanometer. Posterior tibial and dorsal pedal peripheral pulses were palpated. Right and left systolic ankle pressures ``-ere measured supine with the same syphygmomanometer and Doppler probe, detecting flow in the posterior tibial artery, ``-hen possible, or the dorsal pedal artery. The right and left ABIs ``'ere calculated as measures of lower limb arterial stenosis according to the formula: ABI = SBP-1Foot + SBP-2Foot/SBP-1Arm + SBP-2Arm, where SBP1Foot and SBP-2Foot are the ankle systolic blood pressures

American Hearl Journal Volume 134, Number 5, Parl I

taken as a repeat measure of the right or left side. Repeat systolic blood pressure measurements taken from the arm with the higher pressure were used to calculate both right and left ABIs. The minimum ABI was used in the analysis as an indicator of unilateral disease.

Determination of hemostafic factors Blood was collected, after the first 2 to 3 ml ``'ere discarded, into tubes containing 3.8% sodium citrate in a 9:1 blood/anticoagulant ratio. Sampling was performed after a 10-mint, re rest period. Citrated blood was kept on ice and processed within 30 minutes of drawing. Plasma was centrifuged with a plasma separator to remove platelets, aliquoted, and frozen at -70 ~ C until assayed. Assays were performed in duplicate with the same lot numbers of reagents used for all samples. Fibrinogen was assayed according to the Clauss method by determining the clotting time of dilute plasma containing excess thrombin (Dade Division, Baxter Healthcare Corp.) compared with a standardized fibrinogen preparation (Organon-Teknika). Factor VII coagulant activity (VIIc) was measured with a one-stage clotting assay in facto'r-deficient plasma (Organon-Teknika) with thromboplastin (PT-Fib thromboplastin; Instrumentation Laboratories). PAl-1 activity was determined with a chromogenic assay (Spectrolyse PL PAl; American Diagnostica, Inc.). vWf antigen was assayed by enzyme immunoassay (Asserachrom vWf; Diagnostica Stago). Because of the effect of blood type on vWf antigen, blood type was determined whenever feasible.

Statistical analysis Statistical analyses were performed with SAS software (SAS Institute, Caw, N.C.). Continuous variables were expressed as median and interquartile ranges and were compared with the Wilcoxon rank-sum test. The effect of blood group and sex on vWf and the effects of age and sex on fibrinogen and PAI-1 were evaluated according to analysis of variance and analysis of covariance. Percentages ,,'ere compared with Pearson's X2 or Fisher's exact test. All statistical tests were m'o sided. Bivariate correlations were calculated according to Pearson. Logistic regression analysis ``'as used to examine the relations!tip between hemostatic factors and a decreased ABI (minimum of the right or left ABI <0.85) versus normal ABI (minimum ABI >1.0), considering systolic blood pressure, diastolic blood pressure, body mass index, and hip/waist ratio as possible covariates. Age, sex, diabetes, and current smoking status were included in the logistic regression models to control confounding. The likelihood ratio test ``'as used to fit the logistic regression model. When examining the relationship between PAI-1 and ABI, time of draw was included in the models. Odds ratios and 95% confidence intervals vcere computed based on the parametric estimates and standard errors of the coefficients in the univariate and multivariate regression models. Multiple linear regression

979

980

Philipp elal.

Table I. Participant characteristics Charateristic Age (yr); median (interquartile range) Male* Female* White Black Other Current smokers* Diabetes* Aspirin use Blood type A, AB, B O Unknown

Table II. Clinical history of ADMIT participants ADMIT (n = 46)

Control (n = 76)

68 (61-73) 37 (80%) 9 {20%) 42 (91%) 4 {9%) 0 (0%) 13 {28%) 11 (24%) 1 [2%)

67 {64-73) 26 (34%) 50 [66%) 74 (97%) 0 [0%) 2 (3%) 2 [3%) 1 (1%) 4 [5%)

19 {41%) 14 (31%) 13 (28%)

34 [45%) 19 (25%) 23 (30%)

Cardiovascular Chest pain Myocardial infarction Coronary artery bypass graft Perculaneous translum~nal coronary angioplasty Cerebrovascular Transient ischemic attack Carotid endarterectomy Stroke Peripheral arterial disease Claudication Angioplasty Abdominal aneurysm Arterial surgery Amputation

19 (41%) 7 (15%) 12 (26%) 12 (26%) 5 (11%] 8 (17%) 3 [7%) 5 (11%) 3 (7%) 38 (83%) 37 (80%) 8 (17%) 2 (4%) 12 (26%) 1 (2%)

*p < 0.05, chi-scluore.

Table III. Risk factors of peripheral arterial disease* ADMIT (n 46)

Risk factor Hemostatic factors Factor VIIc (%) vWf (%) Fibrinogen (mg/dl) PAl-1 (units/ml)t Other risk factors Systolic blood pressure (ram Hg) Diastolic blood pressure (ram Hg) Cholesterol (mg/dl) Body mass index (kg/m 2) Waist/hip ratio I:

=

Control (n = 76)

p Value

(82-I03) (I12-192) (280-360) { 12.1-26.8}

90 182 270 8.0

(80-103) (156-262) (240-305) (4.4.14.9)

NS 0.009 0.0001 0.04

147(132-160) 83 (74.88) 222 (202-243) 28.3 (25.6-31.8} 0.99 (0.93-I.03}

123 71 218 26.9 0.90

(115-133) (66-79) (188.247) (23.5-28.3) (0.84-0.97)

0.0001 0.0001 NS 0.02 0.0001

92 167 335 14.6

Data are mediansand interquortileranges. NS, Not slgnificonl. * Groupswere comparedwiththeWilcoxanrank.sumlesl. 1pAl-1samplesweredrawnct or before 11AM;controlgroup,n = 31; ADMITgroup,n = 40. tForwaist/hip ratio,n = 73, controlgroup9

analysis was also performed with ABI, fibrinogen, PAI-1, vWf, and factor VIIc as the dependent variable, respectively.

Results C l i n i c a l characteristics

From O c t o b e r 1993 to N o v e m b e r 1994, 46 subjects (37 m e n a n d nine w o m e n ) w e r e recruited as ADI',IIT participants a n d 76 subjects (26 m e n a n d 50 w o m e n ) w e r e recruited as control subjects. The participant characteristics are s h o w n in Table I a n d clinical history of atherosclerotic disease in ADMIT participants is s h o w n in Table II. The m e a n ABI o f ADMIT subjects ( n = 46) was 0.70 + 0.15 c o m p a r e d with 1.13 + 0.08 in control subjects ( n = 76) (p < 0.0001). Age, race,

b l o o d type, and aspirin use w e r e not significantly different b e t w e e n the two groups. There w e r e m o r e current gmokers in the ADMIT g r o u p c o m p a r e d with the control g r o u p (28% vs 3%; p < 0.05) a n d fewer w o m e n (20% vs 66%; p < 0.05). A l l n i n e w o m e n in the ADMIT g r o u p w e r e p o s t m e n o p a u s a l . Forty-nine of the 50 w o m e n in the control g r o u p w e r e postmenopausal and one woman had unknown m e n o p a u s a l status. O n e ADMIT female participant a n d four female control subjects w e r e taking estrog e n - r e p l a c e m e n t therapy.

Hemostatic variables and nonhemostatic risk factors Both systolic a n d diastolic b l o o d pressures w e r e significantly ( p < 0.0001) higher in ADMIT participants

American HeartJournal November 1997

Philipp erol.

Figure 1

Figure 2

Fibrinogen Levels

PAI-i Levels

500 4O

400

El)

E

E

--~"

300

r"

O) 0 200 c ..Q o--

2O

i

I".

J

.

,

9

9

e

99

e

|

< I"1

]0

T e

0 ADMIT

No C V D

ADMIT

Control

Fibrinogen levels in ADMIT subjects (n = 46), ADMIT subjects without heart or cerebrovascular disease (n = 23), and control subjects (n = 76). Mean values are indicated {--).

Figure3

5OO

450

t.,.,0.~

9

400

Ill

I1 9

. 9

9

9

9

_i'o0

J~

"".

~

u. 150 100 5o 0 0

"

It

I

100

II

30

i

i

1

I

I

I

I

0.2

0,4

0.6

0.8

1

1.2

t.4

1.0

ABI

Fibrinogen levels versus ABI in study subjects (n = 122).

c o m p a r e d with the control g r o u p (Table III). In addition, b o d y m a s s index and w a i s t / h i p ratio were higher in the ADMIT g r o u p c o m p a r e d with the control (p < 0.02 a n d p < 0.0001). Cholesterol levels did not differ b e t w e e n tile two groups. Fibrinogen levels were significantly higher in the ADMIT g r o u p c o m p a r e d with the control g r o u p (331 -+

American Head Journal Volume 134, Number 5, Part 1

.;.. 9

No C V D

Jl |

Control

PAl-1 activity in ADMIT subjects (n = 40}, ADMIT subjects without heart or cerebrovascular disease (n = 19), and control subjects (n = 31 ). Values are shown for blood samples drawn before 11 AM. Mean values are indicated (--).

52 mg/dl vs 273 -+ 46 mg/dl; p < 0.0001). W h e n ADMIT subjects with a history o f cardiovascular or cerebrovascular disease were excluded, the remaining subjects still had significantly higher fibrinogen levels c o m p a r e d with control subjects (325 + 55 mg/dl; p < 0.0002) (Fig. 1). PAI-1 levels were also higher in the ADMIT c o m p a r e d with the control group (18.0 _+ 10.1 units/ml vs 10.2 + 7.2 units/ml; p < 0.0001). Because time o f d r a w has b e e n s h o w n to affect PAI-1 levels, with higher PAI-1 levels found in tile morning, 22 the analysis was repeated with PAI-1 samples drawn before 11 a_~t. Similar results w e r e o b s e r v e d (18.7 + 10.0 units/ml vs 13.5 + 8.9 units/ml; p < 0.04). PAI-1 activity was also higher in the ADMIT g r o u p without a history of cardiovascular o r cerebrovascular disease (18.6 + 10.1 units/nil vs 13.5 + 8.9 units/ml), although the differences w e r e not statistically significant (Fig. 2). No differences in vWf antigen or factor \ q l c levels were o b s e r v e d b e t w e e n the two groups. Systolic blood pressure, diastolic blood pressure, and fibrinogen levels were inversely correlated with ABI across the entire cohort. Other correlations between hemostatic and nonhemostatic variables include fibrinogen with factor VIIc, systolic blood pressure and diastolic b l o o d pressure, a n d PAI-1 with b o d y mass index and age (inverse correlation).

981

982

Philippetal.

Table IV, Logistic regressionsof risk factors on peripheral arterial disease Odds ratio (95% Cl) Risk factor Fibrinogen (+ 10%) Factor Vllc (+ 10%) vWf (+ 10%} PAl-1 (+10 units/ml) Systolic blood pressure (+10 mm Hg) Diastolic blood pressure (+10 mm Hg) Body mass index (+1 unit) Waist/hip (+1 unit)

Unadjusted

Adjusted"

1.3(1.2-1.4} 1.1(0.91-1.3) 0.95(0.9GI.00) 1.8(1.1-3.I)

1.3(1.1-1.6)

2.6(1.8-3.8) 4.1(2.3-Z3)

3.8 (2.0-Zl)

1.1(I.GI.2) 16.6(0.42~4.9)

CI, Confidence interval. *Analysis adiusted for age, sex, diabeles, and current smoking status.

To assess the importance of hemostatic factors on the severity o~ disease as measured by ABI, potential hemostatic risk factors were evaluated according to multiple logistic regression analysis (Table IV). Univariate analysis showed that increases in fibrinogen, PAl-l, systolic and diastolic blood pressures, body mass index, and smoking, in addition to male sex and diabetes, were associated with lower limb atherosclerotic narrowing. However, multivariate analysis, controlling for age, diabetes, current smoking, and sex, sbowed that only fibrinogen and systolic blood pressure were independent risk factors of peripheral arterial disease (Table IV). There was no threshold level of fibrinogen associated with an abnormal ABI (Fig. 3). Similar results were obtained when subjects with cardiovascular or cerebrovascular disease were omitted from the analysis. Current smoking became a significant risk factor, in addition to fibrinogen and systolic blood pressure, when the number of variables was reduced by omitting sex and diabetes from the analysis, blultiple linear regression analysis with ABI as the dependent variable also yielded systolic blood pressure and fibrinogen as the major explanatory variables for ABI (data not shown). Because fibrinogen correlated with factor Vllc, fibrinogen was replaced with factor VIIc, and the contribution of factor VIIc to the logistic regression model containing systolic blood pressure was reevaluated. In the absence of fibrinogen, factor VIIc still could not explain ABI status. However, when systolic blood pressure was also omitted from the model, the odds ratio for factor VIIc was 1.3 (95% confidence interval 1.0 to 1:7). Adding either fibrinogen or systolic blood pressure to this model eliminated the association between factor

VIIc and an abnormal ABI. Muhiple linear regression with fibrinogen as the dependent variable confirmed that ABI status and factor VIIc were independently associated with fibrinogen.

Discussion There is increasing evidence for a role of hemostasis in the development of atherosclerosis and an interaction of hemostatic variables with established nonhemostatic risk factors in cardiovascular disease. 1,2,23-25 However, previous studies in peripheral vascular disease with control groups have been limited. 6,26 This study was designed to measure fibrinogen, vWf, PAl-l, and factor VIIc in non-acutely ill subjects and to examine tile relationship of these hemostatic variables with the presence of lower limb arterial occlusive disease, with ABI used as a measure of lower extremity arterial stenosis and taking into account nonhemostatic risk factors. The ABI has been shown to be decreased significantly in subjects with angiographically defined stenosis or occlusion of major arteries to the lower extremities, including aortoiliac obstruction, femoropopliteal obstruction, and combined aortoiliac and femoropopliteal obstruction. 2~ Previous studies have shown that an ABI -<0.9 has a high specificity (>95%) for detecting arteriographically defined stenosis. 29 To reduce the possibility of subjects with a normal ABI having undetected peripheral vascular disease, control subjects for this study were required to have an ABI of >1.0 and no history of claudication or peripheral vascular, cardiovascular, or cerebrovascular disease. According to these criteria, only one diabetic patient was included in the control group.

American Heart Journal November 1997

Philipp elal.

Current smoking was significantly more prevalent in ADMIT subjects compared with control subjects, as has been reported in other studies.6, 26 On multivariate analysis, the association of smoking with ABI was significant when the number of variables analyzed was reduced, possibly reflecting the size of the study cohort. Both current smoking and previous smoking were significantly associated with the ABI in multivariate analysis of hemostatic and rheologic variables in the Edinburgh Artery Study. 6 The major finding of this study supports the premise that fibrinogen is an important independent risk factor for lower limb arterial stenosis. The data show that increases in fibrinogen Concentration, even within the normal range, increase the risk for peripheral yascular disease irrespective of a previous history of cardiovascular or cerebrovascular disease. In addmon no threshold level of fibrinogen was found to i~crease the risk of an abnormal ABI. Only one other study with a comparison group, the Edinburgh :Artery Study, has examined fibrinogen as a risk factor in peripheral arterial disease. The investigators showed that fibrinogen and blood viscosity were independently associated with peripheral arterial narrowing. 6 However, the analysis did not consider peripheral vascular disease unaccompanied by coronary or cerebrovascular disease. 6,26 The demonstration of fibrinogen as an important independent risk factor in symptomatic and asymptomatic peripheral vascular disease parallels the more extensive data on fibrinogen as a risk factor in coronary artery disease.l,3-5, 28 The higher PAI-1 levels observed in subjects with a decreased ABI suggest that deficient fibrinolytic capacity is also associated with peripheral arterial stenosis. Elevated concentrations of PAI-1 have been inlplicated in coronary thrombosis and atherogenesis. 2,23 Although the relationship of increased levels of PAI-1 to the extent of vessel narrowing has been poorly defined in coronary artery disease, elevated PAI-1 activity has been shown to be a predicator of recurrent myocardial infarction 2 and an important detem-tinant of reperfusion after thrombolylic therapy in acute myocardial infarction. 29 In survivors of myocardial infarction, elevated PAI-1 activity has also been associated with increases in body mass index, as we obse~'ed in our subjects with peripheral vascular disease, and trigbveride levels. 3~ In vitro studies suggest that a pathogenic link between lipoprotein metabolism and the regulation of expression of PAI-1 may O c c u r , 31

American Heart Journal VoTume 134, Number& Part I

Despite the premise that endothelial dysfunction is associated with increases in vWf levels, we did not observe differences in vWf antigen levels between the ADMIT and control cohorts, even though the pathogenesis of peripheral vascular disease includes endothelial damage. These data are in contrast to previous observations 26 and may reflect subject numbers, severity of disease, or the similarity in blood types bem'een our two subject groups. Factor VII levels were not significantly different between ADMIT and control subjects but did correlate with cholesterol and fibrinogen levels. Increases in factor VII levels have been reported in coronary artery disease in the Northwick Park Heart Studyt but not in peripheral vascular disease. A weak correlation (0.161; p < 0.01) be~'een factor VII levels and anlde/ama systolic blood pressure has been reported in the Progetto Lombardo Atero Trombosi study,32 which exanfined subjects with documented cardiovascular disease but had no control subjects. We observed a correlation bem'een factor VII levels and cholesterol levels that is consistent with findings in normal adults and pregnant women 33 and in rabbits fed a high-cholesterol diet. 34 The selection criteria of the study, in which potential ADMIT subjects with very high low-density lipoprotein cholesterol or triglyceride levels were excluded from entry, may at least partially explain the absence of increased factor VII levels observed in the ADMIT population. Our results highlight the importance of fibrinogen and PAI-1 activity in peripheral vascular disease and extend the observations noted in ischemic heart disease. The functional significance and pathogenesis of these hemostatic abnormalities in peripheral vascular disease have yet to be determined, and the effects of pharmacologically reducing these variables, particularly fibrinogen, on progressive peripheral vascular disease should prove a fruitful area for future investigation. ~'e thalzk Jea,znelle Print, Maria Propst, and Bridget Troccoli for technical assistance, Robin Schwartz for mamtscript preparation, atul the research staff of the Cardiovascular Diseases and Hypertension Clinical Trials Section.

References 1. Meade TW, Mellows S, Bronzovic M, Miller GJ, Chaakrabarfi RR, North WRS, et al. Hemostatic function and ischaernic heart disease: principal results of lhe Nar|hwick Park Heart Study. lancet 1986;2:533-7. 2. Harnsten A, de Faire U, Walldius G, Dahlen G, Szarnosi A, Landau C, el al. Plasminogen activator inhibitor in plasma: risk faclor for recurrenl myocardial infarction. Lancet 1987;2:3-9. 3. Kannel WB, Wolf PA, Castelli WP, D'Agostino RB. Fibrinogen and risk

983

984

Philipp et al.

of cardiovascular disease: the Framingham study.JAMA 1987;258:1183-6. 4. ErnstE, Resch KL Fibrinogen as a cardiovascular risk factor: a metaanalysis and review of lhe literature.Ann Intern Med 1993;118:95663. 5. Wilheelmsen [, Svardsudd K, Korsan-BengtsenK, [arsson B, Welin [, Tibbhn G. Fibrinogen as a risk factor for stroke and myocardial infarction. N EnglJ Med 1984;311:501-5. 6. Lowe GDO, Fowkes FGR, Dawes J, Donnan PT, Lennie SE, Housley E. Blood viscosity, hbrinogen, and activation of coagulation and leukocyles in peripheral arterial disease and the normal population in the Edinburgh Artery Study9Circulation 1993;87:1915-20. 7. Yarnell .IW, Baker IA, Sweetnam PM, Bainton D, O'Brien JR, Whitehead P.I,el al. Fibrinogen, viscosily, and white blood cell count are major risk foclors for ischemic heart disease: the Caerphilly and Speedwell Collaborative Heart Disease Studies. Circulation 1991;83:83644. 8. Kostis .lB, Baughman J, Kuo PT. Association of recurrent myocardial infarction with hemostatic foclors: a prospective study. Chest 1982;81:571-5. 9. Margulis I", David M, Maor N. The van Wi[lebrand factor in myocari dial infarction and unstable angina: a kinetic study. Thromb Haemost 1986;55:3668. 10. Carvalh9 de Sousa J, Azevedo.I, Soria C, Barros F, Ribeiro C, Parreira F, et al. Factor VII hyperactivity in acute myocardial thrombosis:a relation to the coagulation activation. Thromb Res 1988;51:165-73. 11. Smith EB, Keen GA, Granl A, Stirk C. Fate of hbrinogen in human arterial intima. Arleriosderosis 1990; 10:263-75. 12. Bini A, Fenoglio JJ, Mesa-Tejada R, Kudryk B, Kaplan KI.. Identificatibn and distribution of fibrinogen, fibrin, and fibrin degradation products in atherosclerosis.Arleriosclerosis 1989;9:109-21. 13. Schneiderman J, Sawdey MS, Keeton MR, Bordin GM, Bernstein El:, Dilley RB, et ah Increased type 1 plasminogen activator inhibitor gene expression in atherosderotic human arteries. Proc Natl Acad Sci USA 1992;89:6998-7002. 14. Krishnamurli C, Barr CF, Hassetl MA, Young GD, Alving B. PIasminogen activator inhibitor: a regulator of ancrod induced fibrin deposition in rabbils. Blood 1987;69:798-803. 15. Levi M, Biemond BJ, Van Zonneveld AJ, Ten Cate JW, Pannekoek H. Inhibition of plasminogen activator inhibitor 1 activity resultsin promotion of endogenous thrombolysis and inhibition of lhrombus extension in models of experimental thrombosis9Circulation 1992;85:305-12. 16. Nemerson Y. Tissuefactor and hemostasis. Blood 1988;71:1-8. 1Z Weiss HJ, TurittoVT, Baumgarlner HR, Nemerson Y, Hoffman 1". Evidence for the presence of tissue factor activity on subendothelium. Blood 1989;73:968-75. 18. Badimon [, Badimon ]J, Turitto VT, FusterV. Platelet deposilion in van Willebrand factor deficient vessel wall. J Lob Clin Med 1987; 110:634-4Z 19. FusterV, Bowie EJW, LewisJC, Fass DN, Owen CA, Brown AL Reistanceto arteriosclerosis in pigs with van Willebrand's disease:

spontaneous and high cholesterol diel-induced arteriosclerosis.J Clin Invest 1978;61:722-31. 20. Bernstein EF, Fronek A. Current status of noninv0sive tests in the diagnosis of peripheral arterial disease. Surg Clin North Am 1982;63:473-8Z 21. Fowkes FGR, Allan PL,Tsampoulas C, Smith FB, Donna PT.Validity of duplex scanning in the detection of peripheral arterial disease in the general population. EurJ Vasc Surg 1992;6:31-5. 22. Eliasson M, Evrin PE, tundblad D, Asplund K, Ranby M. Influence of gender, age and sampling time on plasma fibrinolytic variables and fibrinogen, a population study. Fibrinolysis 1993;7:1-8. 23. ]uhan-Vague 1,Alessi MC. Plasminogen activator inhibitor 1 and atherothrombosis.Thromb Haemost 1993;70:138-43. 24. Daae LN, Kierulf P, tandoas S, Urda 'P. Cardiovascular risk factors: inleractive effects of lipids, coaguloti6n and fibrinolysis. Scand J Clin Lab Invest 1993;53:19-2Z 25. Badlmon L, Bad,man JJ, Chesebro J~l, FuslerV. van WiIlebrand factor and cardiovascular disease. Thromib Haemost 1993;70:111-8. 26. Smith FB, Lowe GDO, Fowkes FGR, RumleyA, RumleyAG, Donnan PT, el al. Smoking, haemostatic factors and lipid peroxides in a population case control study of peripheral arlerial disease. Atherosderosis 1993;102:15562. 2Z Carler SA. Indirect systolic pressuresand pulse waves in arterial occlusive disease of the lower extremities.Circulation 1968;38:624-37. 28. Thompson SG, Kienast J, Pyke SD, Haverkale F, van de Loo JCW. Hemoslalic factors and lhe risk of myocardial infarction or sudden death in patients wilh angina pectoris. N Engl J Med 1995;332:635-78. 29. Barbash GF, Had H, Rod A, Miller HI, RathS, Har Xahav Y, el al. Correlation of baseline plasminogen activator inhibitor activity with patency of lhe infarct related arlery after lhrombolytic lherapy in acute myocardial infarction. AmJ Cardio11989;64:1231-5. 30. Gray RP, Mohamed-Ali V, PattersonD, Yudkin JS. Determinantsof plasminogen activator inhibitor-1 activity in survivorsof myocardial infarction. Thromb Haemost 1995;73:261-Z 31. Brown SL, Sobel BE, Fujii S. Attenuation of the synthesisof plasminogen activator inhibilor type 1 by niacin. Circulation 1995;92:767-72. 32. Cortellaro M, Boschetti C, Cofrancesca E, Zanussi C, Catalano M, de Gaetano G, et al. The PLATstudy: a multidisciplinary study of hemostatic function and conventional risk factors in vascular disease patients. Atherosclerosis 1991;90:109-18. 33. Miller GJ, Walter SJ, Stirling Y, Thompson SG, Esnouf MP, Meade TW. Assay of factor VII aclivity by Iwo techniques: evidence for increased conversion of VII to aVIla in hyperlipidaemia, with possible implications for ischaemic heart disease. Br J Haematol 1985;59:249-58. 34. Milropoulos KA, Miller G J, Reeves BEA, Wilkes HC, Cruckshank JK. Factor VII coagulant activity is strongly associated with lhe plasma concentration of large lipoprotein particles in middle-aged men. Atherosderosis 1989;76:203-8.

American Heart Journal November 1997