The Role of Rheological and Haemostatic Factors in Hypertension

REVIEW ARTICLE The role of rheological and haemostatic factors in hypertension

AJ Lee Wolfson Unit for Prevention of Peripheral Vascular Diseases, Department of Public Health Sciences, University of Edinburgh, Teviot Place, Edinburgh EH8 9AG, UK

Research into the pathogenesis of hypertension has, in tors are known to be involved in thrombogenesis, their the past, tended to concentrate on changes in vascular potential role in the pathogenesis of patients with hyper- geometry, cardiac output and blood volume. However, tension has received increasing attention. The purposes patients with hypertension are known to be at an of this review are to update previous reviews and to increased risk of both coronary heart disease and cere- summarise the relationship between hypertension, its brovascular disease; complications which are throm- complications and treatment and a range of haemorheo- botic in origin. Since rheological and haemostatic fac- logical factors.

Keywords: hypertension; rheology; haemostatic factors

Introduction genesis and ischaemia, promoting local hyperviscosity, the dependence of blood viscosity on flow Hypertension is recognised as a major risk factor for conditions suggests that abnormalities in haemo- the development of myocardial infarction and rheological factors may be involved in the compli- 1,2 stroke. Whereas advances in pharmacology have cations of hypertension, as well as in the determi- led to increasingly effective antihypertensive drug nation of BP itself.4 The complications of therapy, the pathological mechanisms of hyperten- hypertension, such as myocardial infarction and sion are still poorly understood. Systemic arterial stroke have a thrombotic basis. The factors generally blood pressure (BP) depends on both cardiac output accepted to be important in thrombus formation are and on total peripheral resistance. In turn, periph- known as Virchow’s triad:5 abnormalities of the ves- eral resistance is affected by the diameter of the sel wall; the composition of the blood (platelets and arterioles and capillaries and also by blood viscosity the coagulation and fibrinolytic systems) and the (Figure 1). flow of blood in the vessel. The purposes of this One of the most important functional rheological review are to update previous reviews4,6,7 and to properties of blood is its resistance to flow or vis- summarise the relationships between hypertension, 3 cosity. The flow streamlines of blood are disturbed its complications and treatment, and a range of rheo- by large plasma proteins and also by the erythro- logical and haemostatic factors, which are increas- cytes in the plasma. At high shear rates, blood vis- ingly recognised as determinants of cardiovascular cosity is low. However, since conditions of low events.8 shear rates arise at sites of atherogenesis, thrombo- In recent years, it has become evident that rheological and haemostatic factors are both primary and secondary predictors of cardiovascular events. Mounting evidence of their prognostic value raises the possibility that they are not merely markers or consequences of atherothrombotic disease, but may also contribute to its pathogenesis. Since the primary factors governing blood viscosity are the haematocrit, plasma viscosity and red cell deformability and aggregation, evidence relating each of these to hypertension will be reviewed in turn. This is followed by brief summaries of those studies which have examined (a) the roles of haemostatic factors in hypertension and (b) the effects of antihypertensive therapy on rheological and haemostatic factors. Figure 1 Haemorheological determinants of blood pressure Haematocrit Correspondence: Dr AJ Lee Table 1 shows that many of the reported studies Received 28 June 1997, revised and accepted 1 September 1997 have confirmed an increased haematocrit among Role of rheological and haemostatic factors AJ Lee 768 Table 1 Mean (s.e.) levels of haematocrit (%) in case-control ised by abnormalities in the haematocrit and an studies of hypertensives and normotensives increased incidence of hypertension. Relative polycythemia is a condition in which patients have an Reference Definition of Cases Controls P-value cases elevated haematocrit despite a normal or decreased total red cell mass. Among subjects with this con- Tibblin et al, 14 M 43.3 ± 40.5 ± Ͻ0.01 dition, two groups can be distinguished: patients 19669 Age 50 0.5 0.7 who have a diminished plasma volume as a result BP у150/у110 n = 14 n = 11 of dehydration and those others who are generally Longhini et al, 13 M, 12 F 43.5 ± 45.6 ± Ͻ0.05 asymptomatic middle-aged men with a history of 198610 Age 22–58 0.6 0.8 hypertension and/or excessive smoking. In 1905, Mean DBP 107.6 n = 25 n = 20 Gaisbock reported that a number of hypertensive ± 2.6 patients had plethora and an elevated red cell count, Puniyani et al, 13 M 43.3 ± 43.3 ± NS but no enlarged spleen,21 a condition now referred 198711 BP Ͻ140/98 1.8 0.9 to as Gaisbock syndrome.22 A later study on patients = = controlled n 13 n 35 with polycythemia led Lawrence and Berlin23 to Letcher et al, 27 M, 22 F 44.0 ± 41.5 ± Ͻ0.005 identify patients with a normal red cell volume, but 198112 Age 45 ± 1.6 0.6 0.4 a reduced plasma volume. Although most of these Ͼ = = Mean AP 107 n 49 n 49 subjects were hypertensive, all exhibited tense and Devereux et al,21M,3F45±0.6 42 ± 0.3 Ͻ0.02 anxious behaviour and so the term ‘stress polycy- 198413 Age 23–63 n = 24 n = 13 Ͼ themia’ was introduced. In polycythemia ruba vera, BP 140/90 red cell volume is increased, which increases blood Kobrin et al,43M45±0.5 43 ± 0.5 Ͻ0.05 viscosity. In all of these disorders, the raised haema- 14 ± = = 1984 Age 41 1 n 43 n 28 tocrit is likely to increase the risk of arterial and DBP Ͼ100 venous thromboembolic events. Bentivoglio et al, 19 M, 10 F 45 42 not In a study of malignant hypertensives, character- 15 ± = = 1985 Age 41.4 1.1 n 29 n 21 given ised by very high BP and changes in retinopathy,24 BP 163.4/102.9 a significantly lower haematocrit among the disease Leschke et al, 27 M, 8 F 45.9 ± 42.3 ± Ͻ0.01 group was found, coupled with increased levels of 16 ± 1990 Age 56 1.4 0.6 1.1 fibrinogen, factor VIIIc and fibrin D-dimer. The SBP Ͼ160 n = 35 n = 10 lower haematocrit was due to the anaemia of renal Sandhagen et al, 16 M, 6 F 47.3 ± 0.7 45.4 ± Ͻ0.05 disease. When these results were compared to a 199017 Mean age 44.3 n = 22 0.4 DBP у105 n = 22 group who did not have the retinal changes of malignant hypertension, but who were matched for age, 18 ± ± Vaya´ et al, 1992 9 M, 12 F 43.1 0.6 42.2 NS sex, smoking habit, renal pathology and levels of Age 28–52 n = 21 0.5 Mean SBP n = 60 serum creatinine and BP, similar blood abnormali- 160 ± 1.3 ties were noted. Since most of the haemostatic variables were significantly correlated with serum creat- Siebers et al, 14 M 46 43 Ͻ0.01 199419 mean age 39 n = 14 n = 14 inine, it is likely that such abnormalities may have DBP 85–95 resulted from renal failure. This is an alternative hypothesis to that proposed by Gavras et al25 who Blood pressure values are mm Hg. suggested that hypertension activates coagulation which leads to intrarenal fibrin deposition and renal failure. It is also possible that renal damage might hypertensives compared with a control group,9–19 then activate haemostasis, promoting further renal although most have been based on a relatively small damage and setting up a vicious circle. number of subjects. In contrast, one recent paper by Table 2 shows that significant correlations Vaya´ et al18 noted no difference in haematocrit level between BP and haematocrit have been reported in when comparing a group of 21 hypertensives with a range of population studies,26–33 although this does a group of controls, which agreed with some earlier not seem to be true among hypertensive patients,34 reports.11,20 An increased haematocrit among hyper- possibly due to the effects of treatment and/or renal tensives could be due to either an increase in red impairment. Most of the epidemiological studies blood cell mass or to a contracted plasma volume. have reported a stronger relationship of haematocrit One study indicated that, compared to a control with diastolic BP (DBP) than with systolic BP group of 24 subjects, 45 patients with mild hyper- (SBP).27–30,32 This is consistent with the contribution tension had similar haematocrit, red blood cell mass of the haematocrit, and hence blood viscosity, to and plasma volume levels. In contrast, 43 other sub- peripheral vascular resistance, which is a determi- jects with moderate to severe hypertension had a nant of DBP.12 There is no agreement as to whether significantly increased haematocrit, which was attri- the relationship between BP and haematocrit is buted solely to a decreased plasma volume, since stronger for men than for women. red blood cell mass was normal.14 In another study,9 compared with normotensives, hypertensives Plasma viscosity showed a decreased plasma volume during exercise with resulting increases in haematocrit, plasma pro- Table 3 shows that increases in plasma viscosity tein and blood viscosity. among hypertensives have been confirmed in vari- There are a range of disorders which are character- ous case-control studies.9,12,16,17,20,35 Letcher et al12 Role of rheological and haemostatic factors AJ Lee 769 Table 2 Correlations between blood pressure (mm Hg) and both the haematocrit (%) and plasma viscosity (mPa.s) in epidemiological studies

Reference Population Haematocrit Plasma viscosity

Carter et al, 198326 8006 M DBP r = 0.23 Age 45–60+ Lowe et al, 198827 668 M, 592 F DBP M r = 0.20; F r = 0.16 DBP M r = 0.12; F r = 0.26 Age 25–64 SBP M r = 0.07; F r = 0.14 SBP M r = 0.13; F r = 0.22 de Simone et al, 199028 75 M, 53 F DBP r = 0.20 DBP r = 0.24 Age 27–75 SBP r = 0.19 SBP r = 0.21 Lowe et al, 199229 477 M, 438 F DBP M r = 0.11; F r = 0.17 Age 25–64 SBP M r = 0.09; F r = 0.15 Smith et al, 199230 1264 M and F DBP M r = 0.25; F r = 0.20 DBP M r = 0.25; F r = 0.27 Age 25–64 SBP M r = 0.16; F r = 0.20 SBP M r = 0.23; F r = 0.27 Bonithon-Kopp et al, 1993*31 637 M, 431 F DBP M r = 0.10; F r = 0.09 Age 45–56 SBP M r = 0.10; F r = 0.07 Fowkes et al, 199332 809 M, 783 F DBP M r = 0.16; F r = 0.10 DBP M r = 0.14; F r = 0.09 Age 55–74 SBP M r = 0.08; F r = 0.01 SBP M r = 0.18; F r = 0.14 Goubali et al, 199533 1067 M, 868 F DBP M and F ␹2 F = 4.61 Age M 52.6 ± 0.4 SBP M and F ␹2 F = 4.61 F 49.3 ± 0.5

*Correlations shown are between change in plasma viscosity and change in blood pressure after a 2-year follow-up.

Table 3 Mean (s.e.) levels of plasma viscosity (mPa.s) in case- results in plasma-volume contraction and haemo- control studies of hypertensives and normotensives concentration, thereby increasing plasma viscosity. Various mechanisms have been proposed to explain Reference Definition of Cases Controls P-value cases the shift in transcapillary fluid. Firstly, increased transcapillary filtration through inadequate modu- Tibblin et al, 14 M 2.08 ± 1.95 ± NS lation of increased arterial pressure by precapillary 1966*9 Age 50 0.10 0.07 sphincter vessels. Secondly, an elevation in vascular BP у150/у110 n = 14 n = 11 venous tone leading to decreased fluid reabsorption Letcher et al, 27 M, 22 F 1.29 ± 1.24 ± Ͻ0.001 via resistance on the venular side of the capillary 198112 Age 45 ± 1.6 0.01 0.01 bed. Lastly, local synthesis of fibrin degradation pro- Mean AP Ͼ107 n = 49 n = 49 ducts in precapillary vessels due to activated Zannad et al, 13 M, 9 F 1.29 ± 1.20 ± Ͻ0.01 interleukins, thereby increasing the hepatic syn- 198520 Age 31–62 0.02 0.01 thesis of fibrinogen and other acute-phase pro- n = 22 n = 30 teins.37,38 26–33 Koenig et al, 29 M, 10 F M 1.29 M 1.23 M Epidemiological studies have tended to con- 198835 Age 25–64 n = 29 n = 274 Ͻ0.001 firm a relationship between plasma viscosity and the BP у160/95 F 1.24 F 1.23 F NS = = normal range of BP (Table 2). A 2-year follow-up of n 10 n 293 a middle-aged French population showed that Leschke et al, 27 M, 8 F 1.39 ± 1.32 ± Ͻ0.01 change in plasma viscosity was significantly associa- 199016 Age 56 ± 1.4 0.01 0.02 Ͼ = = ted with change in both systolic and diastolic BP in SBP 160 n 35 n 10 men, although this association was only of border- Sandhagen et al, 16 M, 6 F 1.38 ± 1.29 ± Ͻ0.001 line significance with diastolic pressure in women.31 17 1990 Mean age 44.3 0.02 0.01 Furthermore, the association with SBP in men per- DBP у105 n = 22 n = 22 sisted after adjustment for smoking status, cholesterol, gamma GT, haemoglobin and fibrinogen. How- Blood pressure values are mm Hg. *Plasma viscosity at shear rate of 23 inverse secs in a rotational ever, baseline plasma viscosity was not found to be viscometer. a significant predictor of developing hypertension during follow-up in those few subjects (25 men and 24 women) who did so. reported that haematocrit-matched hypertensives showed an increased plasma viscosity which largely accounted for the 5–10% increase in blood viscosity Red cell rigidity (lack of deformability) at high shear rates. The MONICA Augsburg study36 showed a strong independent association of plasma Blood flow through the vascular system is depen- viscosity with hypertension in men only, after strati- dent on the ability of the red blood cells to adapt fication for confounding factors. Increased plasma the most suitable configuration for passage through viscosity in arterial hypertension may result from a the capillaries and, when the diameter of the vessel reduction in intravascular fluid volume by a transca- is very small, it is necessary for the red cell to pillary shift of fluid into the interstitial space. An deform during passage through. The deformability elevated capillary hydrostatic filtration pressure of the red blood cells depends mainly on the fluidity Role of rheological and haemostatic factors AJ Lee 770 of the cell membrane, the viscosity of the cytoplasm significantly increased red blood cell aggregation and the cell membrane surface area to cell volume and disaggregation shear rate and shear stress were ratio. Early studies which measured red cell deform- observed in hypertensive subjects. From the epide- ability tended to give conflicting results because of miological perspective, a French population study28 white blood cell and plasma protein contamination. reported that red blood cell aggregation was a more In some studies, the increased blood viscosity important determinant of whole blood viscosity at among hypertensives has not been entirely most shear rates than was plasma viscosity. Results explained by increases in haematocrit or plasma vis- from the second Scottish MONICA Survey29 showed cosity.27,32,39 An early study of 71 hypertensives39 that men had significantly lower red cell aggregation noted a stronger relationship between red cell rigid- levels than women, which could partly be explained ity and SBP than with diastolic pressure. In by their higher haematocrit and lower fibrinogen addition, the effect was more pronounced in men levels. In addition, red cell aggregation correlated than in women. A role for red cell rigidity in the positively with both systolic and diastolic BP in aetiology of hypertension was thus suggested, men (r = 0.12 and r = 0.16, respectively), although although the mechanism of action is still unclear. the correlations in women were weaker (r = 0.07 and Reduced filterability of red cells in hypertensives r = 0.03, respectively). However, neither BP showed has been reported.6,18 Another study used red cell an independent association with red cell aggregation fluidity as a measure of erythrocyte deformability, after adjusting for other cardiovascular risk factors and noted lower levels among subjects with hyper- such as age, smoking, body mass index and lipids. tension.17 It was suggested that this reduced fluidity Such studies suggest that enhanced red cell aggre- may be explained by an enlarged metabolic pool of gation under low shear conditions can lead to an free calcium ions in the red blood cells of patients increased blood viscosity. Furthermore, the obser- with essential hypertension.40 This seems plausible vation that the disaggregation shear rate threshold since earlier work had noted a reduction in the num- is raised in hypertensives42 can be interpreted as a ber of calcium binding sites on the inner surface of stronger adhesive force between the red blood cells the erythrocyte membrane in hypertensives.41 How- within rouleaux. Such parameters may be relevant ever, the fluidity of the cell membrane, on which the to an increased risk of vascular complications which deformability of the erythrocyte depends, may be are associated with low flow conditions including influenced by other factors such as the lipids which atherogenesis, thrombogenesis and ischaemia.3 form part of the membrane bilayer. 30 In the Scottish Heart Health Study, it was shown Blood rheology and the left ventricle in that, after adjustment for plasma viscosity, haematocrit and fibrinogen, the residual variation in whole hypertension blood viscosity (relative blood viscosity, which, Hypertension results in increased pressure afterload when measured at high shear rates, is due to erythro- on the left ventricle producing compensatory left cyte rigidity or lack of deformability) showed no sig- ventricular hypertrophy. The presence of left ven- nificant correlation with BP. In contrast, the Edin- tricular hypertrophy among subjects with hyperten- burgh Artery Study of older subjects did find a sion may be important since it may contribute to relationship between relative blood viscosity and their increased risk of sudden death, coronary heart both systolic and diastolic BP, predominantly in disease, arrhythmias, heart failure and stroke. How- men.32 The general message seems to be that ever, whether the abnormalities in blood rheology impaired red cell deformability may possibly also are a determinant of, or a response to, hypertensive contribute to hyperviscosity in hypertension. left ventricular hypertrophy is unclear. It has been suggested that blood viscosity may be important in Red cell aggregation the pathogenesis of left ventricular hypertrophy because of significant correlations with left ventricu- The aggregation of red blood cells is a reversible lar mass.13,43 Furthermore, hypertensive cardiac phenomenon that occurs with macromolecules hypertrophy was reported to be more closely related bridging the membranes of adjacent cells, and is to rheological abnormalities than to the level of BP. influenced by the shearing forces exerted on blood, It has also been suggested that vasoconstriction may the properties of erythrocytes (including concen- induce both haemoconcentration (resulting in tration, deformability and shape) and the concen- increased blood viscosity), and increased peripheral trations, lengths and bridging force of high molecu- resistance as a stimulus to myocardial hypertrophy. lar weight plasma proteins, especially fibrinogen. In contrast, both stroke volume and plasma volume This phenomenon represents an equilibrium have been positively related to left ventricular mass between the macromolecules bridging force, the in hypertensives.44 Since it is still unclear whether electrical charges on the surface of the red blood blood rheology is a cause or effect of hypertensive cells and the shear stress. When the macromolecules cardiac hypertrophy, further studies are warranted. bridging force is increased in pathological conditions, the shear stress needed to break up red blood cell aggregates is increased. Fibrinogen One early case-control study20 noted that subjects Next to haematocrit, fibrinogen is the most with hypertension had significantly higher levels of important determinant of blood viscosity, due to its red cell aggregation when compared to the controls strong effects on both plasma viscosity and red cell (19.8 ± 0.8 vs 14.0 ± 0.6, P Ͻ 0.01). In a later study,42 aggregation. Fibrinogen is also important in haemo- Role of rheological and haemostatic factors AJ Lee 771 stasis and thrombosis.8 The formation of platelet- that fibrinogen does not have a strong independent rich mural thrombi is enhanced through bridging of role in determining BP in the general population. platelets into aggregates by fibrinogen; it is also the Table 5 shows that most of the studies on subjects precursor of fibrin. with hypertension have shown an elevated fibrino- The association between hypertension and hyper- gen level compared with normotensive con- fibrinogenaemia has long been known. Letcher et trols,12,16,18,59–62 although two have noted no differ- al12 reported that in haematocrit-matched hyperten- ence in fibrinogen level between these two sives, the levels of fibrinogen and other globulins are groups.62,63 Fogari et al64 reported that, among male increased and suggested that this may partly hypertensives, there was only a weak correlation account for their higher blood viscosity. The between fibrinogen and BP (systolic r ϭ 0.14 and increase in fibrinogen was too great to arise wholly diastolic r = 0.06) and similarly with family history from a contracted volume, and may be due to a of hypertension (r = 0.18). When these subjects were chronic phase protein reaction, possibly associated further divided into those with and without a family with increased catecholamine secretion which is a history of hypertension, cigarette smoking and lipid feature of hypertension. abnormalities showed the strongest correlations Early epidemiological data relating fibrinogen to with fibrinogen in the family history group. Thus, BP indicated a significant, positive univariate the presence of a family history of hypertension effect,45–49 although weaker correlations have been among hypertensives would appear to enhance the reported by some.27,50–52 More recent population tendency of fibrinogen to cluster with other cardio- studies29,30,32,52–56 have tended to confirm a weak vascular risk factors in hypertensive patients. association (Table 4). There is disagreement, however, whether this association is stronger for men von Willebrand factor than for women.29,30,32,52,55 Multivariate adjustment for other rheological and cardiovascular risk factors Plasma levels of von Willebrand factor (vWF) are either reduced the association to non-significance in used as a marker of endothelial disturbance. It is the Edinburgh Artery Study,32 or reversed the posi- also an important cofactor for platelet adhesion and tive association in the Scottish Heart Health Study.30 aggregation, especially at high shear rates. Table 6 One explanation for this latter result was that it shows the results from several case-control studies reflected shear-induced degradation of fibrinogen57 which have reported vWF levels.62,65,66 Blann et al65 or chronic low-grade intravascular coagulation.24 reported that vWF levels were significantly higher Regression models attempting to assess the inde- in patients with hypertension who did not have con- pendent influence of BP on fibrinogen levels have current vascular disease or risk factors (131 ± 34 shown varying results. In two,52,55 SBP had a weak, IU/dl) compared with controls (90 ± 30 IU/dl). How- but independent effect in women only. A study of hypertensive diabetic Nigerians found that fibrinogen was the strongest predictor of both systolic and Table 5 Mean (s.e.) levels of plasma fibrinogen (g/l) in case- diastolic BP, followed by blood viscosity and body control studies of hypertensives and normotensives mass index.58 In contrast, other population studies Reference Definition of Cases Controls P-value have reported no independent effect of fibrinogen on cases BP.53,54,56 In general, such research seems to indicate Letcher et al, 27M,22F 3.8 ± 0.1 2.8 ± 0.1 Ͻ0.001 198112 Age 45 ± 1.6 n = 49 n = 49 Mean AP Ͼ107 Table 4 Correlation between blood pressure (mm Hg) and plasma fibrinogen (g/l) in epidemiological studies Landin et al, 11 M 2.6 ± 0.1 2.2 ± 0.2 Ͻ0.05 199059 Age 53 ± 1.5 n = 11 n = 11 SBP Ͼ160 or Reference Population Fibrinogen DBP Ͼ95

Lee et al, 199052 4515 M, 4309 F SBP M r = 0.13; F r = 0.01 Leschke et al, 27 M, 8 F 2.91 ± 2.51 ± Ͻ0.01 Age 40–59 199016 Age 56 ± 1.4 0.11 0.08 SBP Ͼ160 n = 35 n = 10 Møller and 439 M SBP not significant in Kristensen, 199153 Age 51 regression analysis Kannel et al, 1274 M + F 3.14 M 2.84 M M 199260 Age 47–79 3.14 F 2.89 F Ͻ0.001 29 = = Lowe et al, 1992 477 M, 438 F DBP M r 0.12; F r 0.14 F Ͻ0.05 Age 25–64 SBP M r = 0.20; F r = 0.20 Vaya´ et al, 199218 9 M, 12 F 2.60 ± 2.35 ± Ͻ0.05 30 = = Smith et al, 1992 1264 M and F DBP M r 0.03; F r 0.07 Age 28–52 0.09 0.06 = = Age 25–64 SBP M r 0.02; F r 0.12 Mean SBP 160 ± n = 21 n = 60 Folsom et al, 1933 M, 2260 F Significant across quartiles 1.3 199354 Age 18–30 of SBP Cushman et al, 4255 M+F 3.28 ± 3.16 ± Ͻ0.0001 = = 199661 Age Ͼ65 0.01 0.01 Fowkes et al, 809 M, 783 F DBP M r 0.23; F r 0.17 Ͼ = = 199332 Age 55–74 SBP M r = 0.22; F r = 0.09 SBP 160 or n 2046 n 2209 DBP Ͼ95 = = Eliasson et al, 776 M, 807 F DBP M r 0.14; F r 0.23 ± ± 199455 Age 25–64 SBP M r = 0.17; F r = 0.32 Lemne and de 75 M 2.82 2.82 NS Faire, 199662 Age 50 ± 0.1 0.07 0.06 de Boever et al, 745 M DBP r = 0.09 DBP 85–94 n = 75 n = 75 199556 Age 35–59 SBP r = 0.12 Blood pressure values are mm Hg. Role of rheological and haemostatic factors AJ Lee 772 Table 6 Mean (s.e.) levels of von Willebrand factor (IU/dl) in Table 7 Mean (s.e.) levels of plasminogen activator inhibitor case-control studies of hypertensives and normotensives (PAI-1) antigen (U/ml) in case-control studies of hypertensives and normotensives Reference Definition of Cases Controls P-value cases Reference Definition of Cases Controls P-value cases Blann et al, 24M,33F 90±4 131 ± 6 Ͻ0.0001 199365 Age 56 ± 1.6 n = 57 n = 38 Landin et al, 11 M 18.5 ± 7.8 ± 1.3 Ͻ0.05 BP 161/96 199059 Age 53 ± 1.5 4.7 n = 11 Ͼ = ± ± SBP 160 or n 11 Andrioli et al, 14 M, 7F 101.2 86.3 NS DBP Ͼ95 199666 Age 51.3 ± 4.9 10.3 5.6 BP Ͼ160/95 n = 21 n = 20 Jansson et al, 45 M, 39 F 17.8 ± 11.7 ± 0.009 73 ± ± ± 1991 Age 52 0.8 1.5 1.1 Lemne and de 75 M 96 4 101 4NS DBP 90–109 n = 84 n = 54 Faire, 199662 Age 50 ± 0.1 n = 75 n = 75 DBP 85–94 Gleerup et al, 12 subjects 9.22 4.21 0.02 199574 DBP 90–95 n = 12 n = 12 Blood pressure values are mm Hg. Lemne and de 75 M 16.3 ± 13.7 ± 0.032 Faire, 199662 Age 50 ± 0.1 1.6 1.5 DBP 85–94 n = 75 n = 75 ever, vWF was not significantly raised in patients whose hypertension had been controlled by drug Blood pressure values are mm Hg. therapy (104 ± 29 IU/dl). In subjects with uncontrolled or controlled hypertension, who additionally had vascular disease or elevated risk factors, levels of vWF were also raised. A significant correlation taken into account. A strong positive association was noted between absolute levels of vWF and both among hypertensives between PAI-1 and the level systolic and diastolic BP among patients with either of both insulin and lipoprotein and a negative controlled or uncontrolled hypertension in the relationship with the rate of glucose disposal have absence of vascular disease or risk factors. This, been shown.59,62,73 In a study of non-obese hyperten- coupled with the finding that concurrent vascular sive men, treatment with metformin lowered BP and disease or risk factors were not associated with any tended to increase the fibrinolytic activity of the significant further rise in vWF over that seen with subjects.75 Similarly, DBP was independently and uncontrolled hypertension alone, may suggest a inversely associated with tPA activity among hyper- ‘plateau’ effect above which there is no additional tensives.73 Epidemiological studies55,67,76,77 have release of vWF from the endothelial cells. In contrast tended to support the clinical findings with strong to these findings, two case-control studies reported correlations between BP and both PAI-1 and t-PA no significant difference in vWF levels between reported (Table 8). Such studies appear to indicate hypertensive patients and normotensives.62,66 Very that both BP and concomitant metabolic disturb- few population studies have examined the associ- ances may mediate the reduced fibrinolytic capacity ation between vWF and BP in any detail, although (high PAI-1 and tPA antigen levels) observed in the two that have, neither found an independent among hypertensives. effect.67,68 While endothelial disturbance may play a role in hypertension, further studies of its relationship to plasma vWF levels are required. Table 8 Correlations between blood pressure (mm Hg) and both tissue plasminogen activator (t-PA) antigen (ng/ml) and plasmin- Tissue plasminogen activator and ogen activator inhibitor (PAI-1) antigen (U/ml) in epidemiolog- plasminogen activator inhibitor ical studies

Fibrinolysis is one defence system against thrombus Reference Population t-PA PAI-1 formation and mainly depends on tissue plasminogen activator (tPA), which converts plasminogen to Sundell et al, 120 M, DBP M r = 0.26; DBP M r =−0.02; the active plasmin on the surface of the thrombus, 198976 121 F F r = 0.47 F r = 0.33 causing thrombolysis. tPA is rapidly inhibited by Age 30– SBP M r = 0.08; SBP M r =−0.13; = = circulating plasminogen activator inhibitor type 1 60 F r 0.46 F r 0.33 (PAI-1). Both plasma tPA antigen (complexed with Iso et al, 199077 136 M ‘Weak correlations’‘Weak correlations’ PAI-1) and PAI-1 have been shown to have predic- Age 34– tive power for myocardial infarction, stroke and 55 cardiovascular death.69–72 Table 7 shows that PAI-1 Lowe et al, 300 M, DBP M r = 0.30; 67 = levels have been reported to be significantly ele- 1994 303 F F r 0.23 Age 25– SBP M r = 0.24; vated in hypertensives compared to normo- 64 F r = 0.34 tensives,59,62,73,74 whereas tPA antigen did not differ Eliasson et al, 776 M, DBP M r =−0.12; DBP M r = 0.22; significantly, although tPA activity did (this can be 55 =− = 73 1994 * 807 F F r 0.01 F r 0.11 explained by the higher PAI-1 levels). However, Age 25– SBP M r = 0.00; SBP M r = 0.09; since there may be an interaction between metabolic 64 F r = 0.02 F r = 0.20 risk factors (such as glucose and lipoproteins) and the fibrinolytic system, such confounders need to be *Correlations refer to t-PA activity not antigen. Role of rheological and haemostatic factors AJ Lee 773 Other haemostatic factors alterations in rheology in one study,84 while others have reported a beneficial effect on platelet acti- One study reported no significant difference in vation and aggregation in hypertensives.87 More plasma beta-thromboglobulin (a marker of platelet recently, Wysocki et al88 reported that, in a group activation) between hypertensives and normoten- of hypertensive men, a single dose of cadralazine (a 66 sives, whilst a significant elevation in malignant hydralazine-related vasodilator) caused vasodila- hypertensives has been shown, possibly due to tation (a decrease in the total peripheral resistance 24 reduced renal clearance. Fibrin D-dimer is a cross- index) and was accompanied by haemodilution (a linked fibrin degradation product and plasma levels reduction in haematocrit and blood viscosity). A of D-dimer give an index of the turnover of cross- dose-dependent improvement in the red cell rigidity linked fibrin. Although not a blood coagulation fac- index after treatment with urapidil has been tor per se, correlations within the population range reported89, whereas compounds such as diuretics87 of BP, together with elevated levels among hyperten- and omega-3 fatty acids90 appear to have no effect 24,63,67,78,79 sives, have been noted and the epidemio- on platelet function and plasma fibrinogen level. logical studies are summarised in Table 9. Factor However, most of the studies referred to above VII, an important cofactor in the initiation of blood have been short-term clinical trials, often crossover coagulation by the tissue factor pathway showed in design, and all were based on a relatively small weak associations with BP in one epidemiological number of hypertensive patients. No information 67 study, but, together with Factor VIII, has been about the long-term effects of therapy on prognosis 63 shown to be significantly raised in hypertensives. is available and hence there is a need for large scale These individual studies suggest that factors intervention studies which look in particular at the involved in blood coagulation, other than fibrinogen effects of treatment on a wide range of haemostatic and vWF, may have a role in the aetiology of hyper- and rheological factors. tension, although more research is obviously warranted. Significance of haemorheology in hypertension Therapeutic intervention As we have seen, essential hypertension is charac- Because blood viscosity appears to be a determinant terised by an increase in blood viscosity. Although of both increased BP, and its complications this can be explained in some studies by an elevated (thrombotic events and left ventricular haematocrit, plasma viscosity or fibrinogen level, hypertrophy), the effects of antihypertensive ther- other mechanisms such as a reduced red cell apy are of interest. Until recently, it appeared that deformability have been proposed. The mechanisms ␤ vasodilators improved blood rheology, -adrenergic for the haemorheological abnormalities in patients blockers had no effect and diuretics had an adverse with hypertension are not clear. Elevated blood vis- 4,6 effect. Table 10 summarises the recent studies cosity tends to increase BP and to decrease organ which have examined the effects of antihypertensive blood flow (which, in the absence of compensatory drugs on blood coagulation and rheology. Antihy- changes, hinders oxygen delivery, creating an pertensive treatment with calcium-channel blockers ischaemic state); it also increases platelet adhesion has shown limited success in improving the flow to the sub-endothelium, thereby possibly facilitating 80,81 properties of the blood in some studies whilst thrombus formation in stenosed arteries. others have found little or no change in blood or The major complications of essential hyperten- plasma viscosity during chronic vasodilatation with sion are ischaemic heart disease and stroke and sev- 82–84 such medication. Treatment with beta-blockers eral epidemiological studies48–50 have related hae- appears to exert beneficial effects on blood vis- mostatic factors and, more recently, rheological 81,85,86 cosity and reductions in measures of platelet factors91–94 to the incidence of both. A report from 83 activation. ACE inhibitors have shown negligible the Caerphilly and Speedwell studies91 showed that fibrinogen, plasma viscosity, white cell count and Table 9 Correlations between blood pressure (mm Hg) and fibrin smoking predicted ischaemic heart disease just as D-dimer (ng/ml) in epidemiological studies well as did a model incorporating cholesterol, BP and body mass index. In support of this, Lowe et Reference Population Fibrin D-dimer al94 noted that, in the Edinburgh Artery Study, whole blood viscosity was as strong a predictor of Giansante et al, 262 M, 254 F Among 71 hypertensives cardiovascular events in an older population as LDL 199463 Age 46.5 ± 0.5 343 ± 17 vs 301 ± 15 in = cholesterol or DBP, and was stronger than smoking. controls, P 0.056 Furthermore, blood viscosity was a stronger predic- Lowe et al, 199467 300 M, 303 F DBP M r =−0.17; tor of stroke than any of the three conventional risk =− Age 25–64 F r 0.09 factors. Hence, the predictive value of haemorheol- SBP M r = 0.15; F r = 0.12 ogical factors cannot be completely accounted for by their correlations with other risk factors. Rheological 79 =− Lee et al, 1995 737 M, 700 F DBP M r 0.05; abnormalities noted among hypertensives may have Age 55–74 F r =−0.15 SBP M r = 0.06; implications for cardiovascular mortality and mor- F r = 0.18 bidity. Indeed, data from the Leigh Study48 showed that hypertensive subjects with a fibrinogen level Blood pressure values are mm Hg. Ͼ3.5g/l had a 12-fold increased risk of cardiovascu- Role of rheological and haemostatic factors AJ Lee 774 Table 10 Anti-hypertensive drugs and their effects on blood coagulation and rheology

Reference Drug name Signiﬁcant effects

Calcium-channel Amlodipine Whole blood viscosity ↓; Plasma viscosity ↓; Erythrocyte ﬂuidity ↑ blockers: Linde et al, 199681 Wysocki et al, 199282 Isradipine No signiﬁcant effects Ding et al, 199483 Verapamil Beta thromboglobulin ↓ Ding et al, 199484 Nifedipine Platelet factor 4 ↓ Beta-blockers: Frishman, 1991 Celiprolol Fibrinogen ↓ (review)85 Ding et al, 199483 Propranolol Beta thromboglobulin ↓ Haenni and Lithell, Atenolol Fibrinogen ↓ 199686 Linde et al, 199681 Metoprolol Whole blood viscosity ↓ ACE inhibitors: Zannad et al, 199387 Lisinopril Erythrocyte aggregation time ↑; Disaggregation shear rate ↓ Ding et al, 199484 Cilazapril Whole blood viscosity ↑ at a low shear rate; Beta thromboglobulin ↑; Platelet factor 4 ↑ Diuretics: Zannad et al, 199387 Hydrochlorothizide Erythrocyte aggregation time ↓; Disaggregation shear rate ↑ Others: Delamaire et al, Urapidil Erythrocyte rigidity index ↑ 199589 Haenni and Lithell, Urapidil Fibrinogen ↓ 199686 Wysocki et al, 199688 Cadralazine Haematocrit ↓; Whole blood viscosity ↓; Plasma viscosity ↓

lar disease than hypertensive individuals with Wissenschaftlichen Medicin. Von Meidinger Sohn: fibrinogen levels Ͻ2.9g/l. Frankfurt, 1856, pp 458–636. 6 Lowe GDO, Forbes CD. Rheology of cardiovascular disease. In: Lowe GDO (ed). Clinical Blood Rheology, Vol- Concluding remarks ume II. CRC Press: Boca Raton, Fl, 1988, pp 113–140. 7 Lip GYH, Beevers DG. Abnormalities of rheology and Rheological and haemostatic factors would appear coagulation in hypertension. J Hum Hypertens 1994; to be implicated in the pathogenesis of hyperten- 8: 693–702. sion, its complications and treatment. However, the 8 Lowe GDO. Haemostatic risk factors for arterial and exact mechanisms through which they act are still venous thrombosis. In: Poller L, Ludlam CA (eds). unclear. Mounting evidence indicates that whole Recent Advances in Blood Coagulation, 7. Churchill blood viscosity and its determinants are elevated Livingstone: Edinburgh, 1997, pp 69–96. among patients with hypertension. Plausible roles 9 Tibblin G, Bergentz SE, Bjure J, Wilhelmsen L. Hema- for erythrocyte deformability and aggregation, endo- tocrit, plasma protein, plasma volume, and viscosity thelial dysfunction, platelets and prothrombotic in early hypertensive disease. Am Heart J 1966; 72: markers have also been discussed. Further clinical 165–176. and epidemiological studies are needed to elucidate 10 Longhini E et al. Hemorheology in hypertension. Clin the role of abnormalities in haemorheological factors Hemorheol 1986; 6: 567–576. 11 Puniyani RR, Annapurna V, Chaturani P, Kale PA. among hypertensives and subsequent implications Haemorheological profile in cases of hypertension. for treatment. Clin Hemorheol 1987; 7: 767–772. 12 Letcher RL et al. Direct relationship between blood Acknowledgement pressure and blood viscosity in normal and hypertensive subjects. Role of fibrinogen concentration. Am J Dr Lee is funded on a programme grant from the Med 1981; 70: 1195–1202. British Heart Foundation. 13 Devereux RB et al. Whole blood viscosity as a determinant of cardiac hypertrophy in systemic hypertension. Am J Cardiol 1984; 54: 592–595. References 14 Kobrin E et al. Stable red cell mass despite contracted 1 MacMahon S et al. Blood pressure, stroke and coron- plasma volume in men with essential hypertension. J ary heart disease. Lancet 1990; 335: 765–774. Lab Clin Med 1984; 104: 11–14. 2 Kannel WB. The clinical heterogeneity of hyperten- 15 Bentivoglio M et al. Blood viscosity as a determinant sion. Am J Hypertens 1991; 4: 283–287. of left ventricular wall thickness and mass in systemic 3 Lowe GDO. Blood rheology in arterial disease. Clin Sci hypertension. Clin Hemorheol 1985; 5: 716 (abstract). 1986; 71: 137–146. 16 Leschke M, Vogt M, Motz W, Strauer BE. Blood rhe- 4 Lowe GDO. Blood rheology in general medicine and ology as a contributing factor in reduced coronary surgery. Ballie` re’s Clin Haematol 1987; 1: 827–861. reserve in systemic hypertension. Am J Cardiol 1990; 5 Virchow R. Phlogose und Thrombose in Gera¨␤system. 65: 56G–59G. In: Virchow R (ed). Gesammelte Abhandlungen zur 17 Sandhagen B, Frithz G, Waern U, Ronquist G. Role of rheological and haemostatic factors AJ Lee 775 Increased whole blood viscosity combined with 39 Dintenfass L, Girolami A. Rigidity of red cells in essen- decreased erythrocyte fluidity in untreated patients tial hypertension. Haemostasis 1978; 7: 298–302. with essential hypertension. J Int Med 1990; 228: 40 Ronquist G, Frithz G. Decreased 45calcium uptake in 623–626. red cells of patients with essential hypertension. Acta 18 Vaya´ A, Martinez M, Garcia J, Aznar J. Hemorheolog- Med Scand 1988; 224: 445–449. ical alterations in mild essential hypertension. 41 Orlov SN, Postnov YV. Ca2+ binding and membrane Thromb Res 1992; 66: 223–229. fluidity in renal and essential hypertension. Clin Sci 19 Siebers RW, Carter JM, Maling TJ. Increase in haema- 1982; 62: 281–284. tocrit in borderline hypertensive men. Clin Exp Pharm 42 Razavian SM, Pino MD, Simon A, Levenson J. Increase Physiol 1994; 21: 401–403. in erythrocyte disaggregation shear stress in hyperten- 20 Zannad F et al. Hemorheological disturbances in sion. Hypertension 1992; 20: 247–252. hypertension. Clin Hemorheol 1985; 5: 716 (abstract). 43 Zannad F et al. Haemorheological abnormalities in 21 Gaisbock F. Die Bedeutung des Blutdruckmessung fu¨ r arterial hypertension and their relation to cardiac die a¨rztlichen Praxis. Dtsch Arch Lin Med 1905; 83: hypertrophy. J Hypertens 1988; 6: 293–297. 363–364. 44 Ganau A et al. Stroke volume and left heart anatomy in 22 El-Yousef MK, Bakewell WE Jr. The Gaisbock syn- relation to plasma volume in essential hypertension. J drome. JAMA 1972; 220: 864 (abstract). Hypertens 1992; 10: 725–732. 23 Lawrence JH, Berlin NI. Relative polycythemia – the 45 Korsan-Bengtsen K, Wilhelmsen L, Tibblin G. Blood polycythemia of stress. Yale J Biol Med 1952; 24: coagulation and fibrinolysis in a random sample of 788 498–501. men 54 years old. Thromb Diathes haemorrh 1972; 28: 24 Isles C et al. Abnormal haemostasis and blood vis- 99–108. cosity in malignant hypertension. Thromb Haemost 46 Wilhelmsen L et al. Fibrinogen as a risk factor for 1984; 52: 253–255. stroke and myocardial infarction. N Engl J Med 1984; 25 Gavras H et al. Abnormalities of coagulation and the 311: 501–505. development of malignant phase hypertension. Kidney 47 Balleisen L et al. Epidemiological study on factor VII, Int 1975; 8: 252–261. factor VIII and fibrinogen in an industrial population – 26 Carter C et al. Hematocrit and the risk of coronary ii. Baseline data on the relation to blood pressure, heart disease: the Honolulu Heart Program. Am Heart blood glucose, uric acid, and lipid fractions. Thromb J 1983; 105: 674–679. Haemost 1985; 54: 721–723. 27 Lowe GDO et al. Cardiovascular risk and 48 Stone MC, Thorp JM. Plasma fibrinogen – a major cor- haemorheology – results from the Scottish Heart onary risk factor. J R Coll Gen Pract 1985; 35: 565–569. Health Study and the MONICA project, Glasgow. Clin 49 Kennel WB, Wolf PA, Castelli WP. D’Agostino RB. Hemorheol 1988; 8: 517–524. Fibrinogen and risk of cardiovascular disease. The Fra- 28 de Simone G et al. Relation of blood viscosity to demo- mingham study. JAMA 1987; 258: 1183–1186. graphic and physiological variables and to cardio- 50 Meade TW et al. Characteristics affecting fibrinolytic vascular risk factors in apparently normal adults. Cir- activity and plasma fibrinogen concentrations. Br Med culation 1990; 81: 107–117. J 1979; 1: 153–156. 29 Lowe GDO et al. Epidemiology of haematocrit, white 51 Iso H et al. Plasma fibrinogen, factor VIIc, factor VIIIc, cell count, red cell aggregation and fibrinogen: the and von Willebrand factor and their relations to car- Glasgow MONICA study. Clin Hemorheol 1992; 12: diovascular disease risk factors. Am J Epidemiol 1989; 535–544. 130: 925–934. 30 Smith WCS, Lowe GDO, Lee AJ, Tunstall-Pedoe H. 52 Lee AJ, Smith WCS, Lowe GDO, Tunstall-Pedoe H. Rheological determinants of blood pressure in a Scot- Plasma fibrinogen and coronary risk factors: the Scot- tish adult population. J Hypertens 1992; 10: 467–472. tish Heart Health Study. J Clin Epidemiol 1990; 43: 31 Bonithon-Kopp C et al. Longitudinal associations 913–919. between plasma viscosity and cardiovascular risk fac- 53 Møller L, Kristensen TS. Plasma fibrinogen and tors in a middle-aged French population. Atheroscler- ischaemic heart disease risk factors. Arterioscler osis 1993; 104: 173–182. Thromb 1991; 11: 344–350. 32 Fowkes FGR et al. The relationship between blood vis- 54 Folsom AR et al. Plasma fibrinogen: levels and cor- cosity and blood pressure in a random sample of the relates in young adults. Am J Epidemiol 1993; 138: population aged 55 to 74 years. Eur Heart J 1993; 14: 1023–1036. 597–601. 55 Eliasson M, Evrin P-E, Lundblad D. Fibrinogen and 33 Goubali A et al. Relation of hematocrit values to coron- fibrinolytic variables in relation to anthropometry, lip- ary heart disease, arterial hypertension and respiratory ids and blood pressure. The Northern Sweden MON- impairment in occupational and population groups of ICA study. J Clin Epidemiol 1994; 47: 513–524. the Athens area. Angiology 1995; 46: 719–725. 56 de Boever E et al. Relation of fibrinogen to lifestyles 34 Lowe GDO et al. Haematocrit in patients attending a and cardiovascular risk factors in a working popu- hypertension clinic. Scot Med J 1985; 30: 168–172. lation. Int J Epidemiol 1995; 24: 915–921. 35 Koenig W, Ernst E, Matrai A. Blood rheology associa- 57 Charm SE, Wong BL. Shear degradation of fibrinogen ted with cardiovascular risk factors and chronic car- in the circulation. Science 1970; 170: 466–468. diovascular diseases: results of an epidemiological 58 Memeh CU. The relationship between body weight cross-sectional study. Angiology 1988; 39: 986–995. and plasma viscosity in hypertensive diabetic Nigeri- 36 Koenig W et al. Is increased plasma viscosity a risk ans. J Hypertens 1990; 8: 711–714. factor for high blood pressure? Angiology 1989; 40: 59 Landin K, Tengborn L, Smith U. Elevated fibrinogen 153–163. and plasminogen activator inhibitor (PAI-1) in hyper- 37 Chrysant SG et al. Pathophysiologic significance of tension are related to metabolic risk factors for cardio- ‘stress’ or relative polycythemia in essential hyperten- vascular disease. J Int Med 1990; 227: 273–278. sion. Am J Cardiol 1976; 37: 1069–1072. 60 Kannel WB, D’Agostino RB, Belanger AJ. Update on 38 Ritchie DG, Fuller GM. Hepatocyte-stimulating factor: fibrinogen as a cardiovascular risk factor. Ann Epide- a monocyte derived acute phase regulatory protein. miol 1992; 2: 457–466. Ann NY Acad Sci 1983; 408: 490–502. 61 Cushman M et al. Association of fibrinogen and coagu- Role of rheological and haemostatic factors AJ Lee 776 lation factors VII and VIII with cardiovascular risk fac- 78 Vaziri ND et al. Coagulation and inhibitory and tors in the elderly. Am J Epidemiol 1996; 143: 665– fibrinolytic proteins in essential hypertension. J Am 676. Soc Nephrol 1993; 4: 222–228. 62 Lemne C, de Faire U. Elevation of plasminogen acti- 79 Lee AJ, Fowkes FGR, Lowe GDO, Rumley A. Determi- vator inhibitor 1 in borderline hypertension is linked nants of fibrin D-dimer in th Edinburgh Artery Study. to concomitant metabolic disturbances. Eur J Clin Arterioscler Thromb Vasc Biol 1995; 15: 1094–1097. Invest 1996; 26: 692–697. 80 Koenig W, Ernst E. The effects of calcium channel 63 Giansante C et al. Fibrinogen, D-dimer and thrombin- blockers on blood fluidity. J Cardiovasc Pharmacol antithrombin complexes in a random population sam- 1990; 6 (Suppl 6): s40–s44. ple: relationships with other cardiovascular risk fac- 81 Linde T et al. Decreased blood viscosity and serum tors. Thromb Haemost 1994; 71: 581–586. levels of erythropoietin after anti-hypertensive treat- 64 Fogari R et al. Associations between plasma fibrinogen ment with amlodipine or metoprolol: results of a levels and cardiovascular risk factors in hypertensive cross-over study. J Hum Hypertens 1996; 10: 199–205. men. J Cardiovasc Risk 1994; 1: 341–345. 82 Wysocki M et al. Flow resistance and its components 65 Blann AD, Naqvi T, Waite M, McCollum CN. von Wil- in hypertensive men treated with the calcium antagon- lebrand factor and endothelial damage in essential ist isradipine. Eur J Clin Pharmacol 1992; 43: 463–468. hypertension. J Hum Hypertens 1993; 7: 107–111. 83 Ding YA, Chou TC, Lin KC. Effects of long-acting pro- 66 Andrioli G et al. Study of platelet adhesion in patients pranolol and verapamil on blood pressure, platelet with uncomplicated hypertension. J Hypertens 1996; function, metabolic and rheological properties in 14: 1215–1221. 67 Lowe GDO, Rumley A, Lee AJ, Tunstall-Pedoe HD. hypertension. J Hum Hypertens 1994; 8: 273–278. Epidemiology of plasma D-dimer, t-PA and Von Wille- 84 Ding YA, Law HW, Chou TC. Comparison of cilazapril brand factor antigens: Glasgow MONICA study. and nifedipine retard on ambulatory blood pressure, Fibrinolysis 1994; 8 (Suppl 1): 19 (abstract). metabolic, rheological and platelet function in hyper- 68 Conlan MG et al. Associations of factor VIII and von tensive patients. J Hum Hypertens 1994; 8: 137–143. Willebrand factor with age, race, sex and risk factors 85 Frishman WH. Clinical perspective on celiprolol: car- for atherosclerosis. The Atherosclerosis Risk in Com- dioprotective potential. Am Heart J 1991; 121: 724– munities (ARIC) study. Thromb Haemost 1993; 70: 729. 380–385. 86 Haenni A, Lithell H. Urapidil treatment decreases 69 Meade TW et al. Haemostatic function and ischaemic plasma fibrinogen concentration in essential hyperten- heart disease: principal results of the Northwick Park sion. Metabolism 1996; 45: 1221–1229. Heart Study. Lancet 1986; I: 533–537. 87 Zannad F et al. Effects of lisinopril and hydrochloro- 70 Hamsten A et al. Plasminogen activator inhibitor in thiazide on platelet function and blood rheology in plasma: risk factor for recurrent myocardial infarction. essential hypertension: a randomly allocated double- Lancet 1987; ii: 3–8. blind study. J Hypertens 1993; 11: 559–564. 71 Juhan-Vague I, Alessi MC. Plasminogen activator 88 Wysocki M et al. Hemorheologic effects of vaso- inhibitor 1 and atherothrombosis. Thromb Haemost dilation in essential hypertension. Angiology 1996; 47: 1993; 70: 138–143. 869–878. 72 Ridker PM et al. Prospective study of endogenous 89 Delamaire M, Bernard MD, Boisseau MR. Hemorheo- tissue plasmaingen activator and risk of stroke. Lancet logy of arterial hypertension: influence of urapidil. 1994; 343: 940–943. Blood Pressure 1995; Suppl 3: 58–61. 73 Jansson J-H, Johansson B, Boman K, Nilsson TK. Hypo- 90 Schmidt EB, Dyerberg J. Omega-3 fatty acids. Current fibrinolysis in patients with hypertension and elevated status in cardiovascular medicine. Drugs 1994; 47: cholesterol. J Int Med 1991; 229: 309–316. 405–424. 74 Gleerup G, Vind J, Winther K. Platelet function and 91 Yarnell JWG et al. Fibrinogen, viscosity, and white fibrinolytic activity during rest and exercise in border- blood cell count are major risk factors for ischaemic line hypertensive patients. Eur J Clin Invest 1995; 25: heart disease. The Caerphilly and Speedwell Collabor- 266–270. ative Heart Disease Studies. Circulation 1991; 83: 75 Landin K, Tengborn L, Smith U. Treating insulin 836–844. resistance in hypertension with metformin reduces 92 Wannamethee G, Perry IJ, Shaper AG. Haematocrit, both blood pressure and metabolic risk factors. J Int hypertension and risk of stroke. J Int Med 1994; 235: Med 1991; 229: 181–187. 163–168. 76 Sundell IB et al. Fibrinolytic variables are related to 93 Wannamethee G, Shaper AG, Whincup PH. Ischaemic age, sex, blood pressure, and body build measure- heart disease: association with the haematocrit in the ments: a cross-sectional study in Norsjo¨, Sweden. J Clin Epidemiol 1989; 42: 719–723. British Regional Heart Study. J Epidemiol Comm Hlth 77 Iso H et al. Hemostatic variables in Japanese and Cau- 1994; 48: 112–118. casian men. Tissue plasminogen activator, antithrom- 94 Lowe GDO et al. Blood viscosity and risk of cardio- bin III and protein C and their relations to coronary vascular events: the Edinburgh Artery Study. Br J risk factors. Am J Epidemiol 1990; 132: 41–46. Haematol 1997; 96: 168–173.