Journal of Human Hypertension (2000) 14, 687–690 2000 Macmillan Publishers Ltd All rights reserved 0950-9240/00 $15.00 www.nature.com/jhh REVIEW ARTICLE Hypertension and the prothrombotic state GYH Lip Haemostasis Thrombosis and Vascular Biology Unit, University Department of Medicine, City Hospital, Birmingham, UK The basic underlying pathophysiological processes related to conventional risk factors, target organ dam- underlying the major complications of hypertension age, complications and long-term prognosis, as well as (that is, heart attacks and strokes) are thrombogenesis different antihypertensive treatments. Further work is and atherogenesis. Indeed, despite the blood vessels needed to examine the mechanisms leading to this being exposed to high pressures in hypertension, the phenomenon, the potential prognostic and treatment complications of hypertension are paradoxically throm- implications, and the possible value of measuring these botic in nature rather than haemorrhagic. The evidence parameters in routine clinical practice. suggests that hypertension appears to confer a Journal of Human Hypertension (2000) 14, 687–690 prothrombotic or hypercoagulable state, which can be Keywords: hypercoagulable; prothrombotic; coagulation; haemorheology; prognosis Introduction Indeed, patients with hypertension are well-recog- nised to demonstrate abnormalities of each of these Hypertension is well-recognised to be an important 1 components of Virchow’s triad, leading to a contributor to heart attacks and stroke. Further- prothrombotic or hypercoagulable state.4 Further- more, effective antihypertensive therapy reduces more, the processes of thrombogenesis and athero- strokes by 30–40%, and coronary artery disease by 2 genesis are intimately related, and many of the basic approximately 25%. Nevertheless the basic under- concepts thrombogenesis can be applied to athero- lying pathophysiological processes underlying both genesis. Importantly, recent improvements in bio- of these major complications of hypertension are chemical techniques have enabled us to quantify dif- thrombogenesis and atherogenesis. Indeed, despite ferent components of both these processes. the blood vessels being exposed to high pressures in hypertension, the complications of hypertension are paradoxically thrombotic in nature rather than Evidence for the prothrombotic state in haemorrhagic. Whilst much attention has been hypertension focused on the renin-angiotensin system, cathechol- Evidence for the hypercoagulable state in hyperten- amines and other neurohormonal mechanisms sion has been extensively reviewed.4–6 Indeed, evi- involved in the pathogenesis of hypertension, the dence from numerous epidemiological,7–9 cross- study of the prothrombotic state in hypertension has sectional10,11 and cohort studies12,13 have reported been relatively neglected. abnormalities in the coagulation and fibrinolytic Over 150 years ago, Virchow postulated that three pathways, as well as in platelets and the endo- features predispose to thrombus formation, that is, thelium. abnormalities in blood flow, blood constituents and the vessel wall.3 Whilst Virchow was referring to venous thrombosis, the concepts can be applied to Relation to conventional risk factors arterial thrombosis. An update of Virchow’s triad for These abnormalities in haemostasis appear to be thrombogenesis for the new millennium can be con- additive to conventional risk factors for cardiovascu- sidered by reference to abnormalities of haemo- lar and cerebrovascular events. For example, in the rheology and turbulence at bifurcations and stenotic ECAT study, high plasma fibrinogen in association regions (that is, ‘abnormal blood flow’), abnormali- with high serum cholesterol was associated with the ties in platelets as well as the coagulation and highest risk for cardiovascular events.14 The inter- fibrinolytic pathways (‘abnormal blood con- action between plasma fibrinogen and cholesterol stituents’) and finally, abnormalities in the endo- 4 levels is also demonstrated in the Leigh Study, thelium (‘abnormal vessel wall’). where the incidence of heart attacks was six times greater in those with high plasma fibrinogen (у3.5 g/l) and cholesterol levels (у6.2 mmol/l), Correspondence: Dr GYH Lip, Haemostasis Thrombosis and when compared to those with low fibrinogen Vascular Biology Unit, University Department of Medicine, City Ͻ 15 Hospital, Birmingham B18 7QH, UK. ( 3.5 g/l) levels. Other risk factors, such as smok- E-mail: [email protected] ing and diabetes, markedly influence the prothrom- Received and accepted 18 March 2000 botic state and are probably additive to the intrinsic Hypertension and the prothrombotic state GYH Lip 688 abnormalities (and cardiovascular risk) seen in merely markers or consequences of atherothrom- hypertensives. botic disease, but may contribute to the pathogen- esis of hypertension and its complications. Indeed Association with target organ damage various indices are predictive of outcome in hyper- tension. The abnormalities in haemostasis in hypertensives For example the Leigh general practice study can be related to target organ damage, such as the reported that hypertensive subjects with plasma presence of left ventricular hypertrophy on echocar- Ͼ 10 fibrinogen levels 3.5 g/l had a 12-fold higher car- diography. The latter is a powerful predictor of diovascular risk than those with plasma fibrinogen cardiovascular events, with an eight-fold increase in levels Ͻ2.9 g/l.15 Blann et al26 suggested that high the risk of stroke and a four-fold increase in the risk von Willebrand factor levels had prognostic value of coronary artery disease. The presence of left ven- in hypertension, being predictive of cardiovascular tricular hypertrophy (LVH) is also an independent disease progression. The study by Agewall et al12 contributor to the risk of stroke in atrial fibril- + 16 found that prothrombin fragment 1 2 and C-reactive lation. Furthermore, high von Willebrand factor protein were independent predictors of major coron- levels, an established index of endothelial damage ary events. Our recent study also suggested that or dysfunction can be related to microalbuminuria patients with hypertension who developed cardio- (defined as the excretion of urine albumin between vascular or cerebrovascular events at 4 years’ follow- 20 and 200 mcg/min), another surrogate manifes- 17 up had higher baseline vWf and fibrin D-dimer lev- tation of hypertensive target organ damage. els compared to those without events, although on The abnormalities in various indices can perhaps Cox multivariate proportional hazards analysis only be related to the degree, and possibly the duration plasma fibrinogen and blood pressure levels of hypertension, and those with mild hypertension emerged as independent predictors.13 or lower blood pressures, and more recent onset The possibility therefore remains that some hypertension (which is usually more difficult to pre- prothrombotic indices, either individually or combi- cisely quantify) may show less abnormalities in the nation, may provide sufficiently high predictive prothrombotic state. For example, patients with Ͼ value for cardiovascular disease and stroke. Further severe hypertension (defined as 160/95 mm Hg) prospective studies on large cohorts would be demonstrate high plasma von Willebrand factor required to confirm this hypothesis. levels10,18 which does not appear to be present in patients with milder elevations of blood pressure.7,8 Although endothelial dysfunction or damage can be Effects of treatment present as a result of hypertension, others have even Treating hypertension may reduce the prothrom- considered that endothelial damage may actually 19 botic state. Indeed, antihypertensive agents with promote hypertension. particular benefits in the hypercoagulable state in hypertensives would be likely to have additional Association with the complications of advantages in reducing stroke and other throm- hypertension boembolic events. For example, treated hyperten- sives demonstrate normal von Willebrand factor The common complications of hypertension can 18 also be related to a prothrombotic state. For levels. However, different drugs may affect the prothrombotic state differently (as reviewed by example, hypertension is a common cause of atrial 6 fibrillation20 and indeed, is additive to the risk of Lee ). For example, drugs such as beta-blockers or stroke and thromboembolism with this arrhyth- calcium antagonists may have favourable haemo- 21 rheological actions. In contrast, diuretics may have mia. Atrial fibrillation per se is also well-recog- 27,28 nised to be associated with abnormalities of haemo- the opposite effect in increasing blood viscosity. stasis and endothelial dysfunction, which are These differences may in part explain some of the altered by cardioversion and antithrombotic ther- differences between different antihypertensive apy, and are independent of underlying aetiology or agents in the reduction of endpoints in some trials structural heart disease.22 of antihypertensive therapy. For example, thiazides Hypertension is an important cause of heart fail- are beneficial in older hypertensives (over beta- 23 blockers and placebo) in reducing stroke and cardiac ure and the evidence also points towards a hyper- 29,30 coagulable state in heart failure.24 Indeed, heart fail- events. In contrast, hypertensive patients on ure is an important contributor to stroke and diuretic therapy have an increased mortality if elec- trocardiographic abnormalities (including