Coagulation factors, activation markers and risk of coronary heart disease: the Northwick Park Heart Studies.

In 1976, Hampton and Mitchell concluded their textbook chapter on thrombosis thus: The study of the epidemiology of arterial and venous disease has provided some indicators of where we might look for the causes of thrombosis, but the mechanisms involved are far from clear.... It is, however, possible that empirical advances may occur in the identification of anti-thrombotic drugs despite our ignorance of the basic nature of the mechanism of thrombosis [1]. This statement emphasizes the gap between evidence from randomized trials of empirical antithrombotic therapies in coronary heart disease (CHD) prevention and treatment, and scientific evaluation of the mechanisms through which activation of platelets and coagulation promote thrombosis in coronary arteries (and indeed, other arteries, veins, and the heart chambers). By 1976, Meade and colleagues had established the first Northwick Park Heart Study (NPHS I) in London, UK, to determine the population distributions and associations with thrombotic risk factors of platelet, coagulation andfibrinolytic variables [2], and in due course their associations with CHD risk [3–5]. Reports fromNPHSI showedassociations betweenCHDrisk and some coagulation factor levels (fibrinogen, factors VII and VIII) and vonWillebrand factor (VWF),whichhave since been confirmed by meta-analyses of prospective studies for fibrinogen [6] and VWF [7]. The associations of factor (F) VII and VIII levels with CHD risk remain to be established with confidence in further studies [8]. Studies in the Netherlands of patients with hemophilia and carriers suggest a dose-dependent relationship between factors VIII and IX and CHD risk [9,10], as do case– control studies ofmyocardial infarction (MI) [11]. Furthermore, ameta-analysis of the two commongenetic thrombophilias (FV Leiden and the prothrombin G20120A mutation) show significant associations with CHD risk [12], albeit 10 times weaker than their associations with risk of venous thrombosis. Such epidemiological studies of coagulation factor levels (and related functional genotypes) have the major limitation that they indicate only the potential for coagulation activation. In order to assess actual coagulation activation, epidemiological studies of coagulation activation markers are required. Such studies face major challenges. Some markers [e.g. fibrinopeptide A (FpA), thrombin-antithrombin (TAT) complexes, prothrombin fragment F1 + 2] are susceptible to artefactual elevation during venepuncture, and have short halflives (minutes) in the circulation, which limit the power of studies to detect transient coagulation activation. Furthermore, sampling of peripheral venous blood may not reflect coagulation activation in the local coronary circulation. The associations of fibrin D-dimer with risk of CHD (and other thrombotic events) [13,14] may be due to its relatively long half-life in the circulation (days) and its lack of susceptibility to venepuncture activation. However, further studies are required to establish with confidence the strength of its association with CHD risk [13,14], and importantly the levels of the antecedent coagulation cascade at which activation is associated with CHD risk. In this issue of the Journal, the late George Miller and his colleagues [15] report a study of coagulation activationmarkers at various levels of the coagulation cascade with CHD risk in the secondNorthwick ParkHeart Study (NPHS II). Because of increasing evidence that some coagulation factors and activation markers are associated with markers of the acute phase protein response such as C-reactive protein (CRP) [16], they also studied the associations of coagulation activation markers with CRP levels. Each coagulation activation marker was associated with CRP levels, a finding consistent with the wellestablished association of coagulation factors and their activation with the inflammatory response [16,17]. However, in a multivariate analysis including conventional CHD risk factors and CRP, a one SD increase in FIX activation peptide was associated with an increase in CHD risk (relative hazard 1.20; 95% CI: 1.00–1.43) and a one SD increase in FVIIa level was associated with a decrease in CHD risk (0.70; 95% CI: 0.58– 0.86). In contrast to a previous report of 104 CHD events from Correspondence: G. D. O. Lowe, University of Glasgow, Royal Infirmary, 10 Alexandra Parade, Glasgow G31 2 ER, UK. Tel.: +44 141 211 5412; fax: +44 141 211 0414; e-mail: g.d.lowe@ clinmed.gla.ac.uk Journal of Thrombosis and Haemostasis, 6: 256–258