Predictive Haemostatic Variables In Cardiovascular Disease

July 20, 2002
8:00 to 12:00
Terrace Room
Boston Park Plaza Hotel

Chairman: L. Iacoviello, Italy
Co-Chairs: M. Cushman, USA; P. Grant, UK; R. Hull, Canada; G. Lowe, UK

The number of attendees of this subcommittee meeting was approximately 100.

Dr. Licia Iacoviello gave an update of the activities of the Subcommittee during the past year. A WEB site of this SubCommittee has been developed at http://www.negrisud.it/ssc and linked with the official ISTH website.

The Registry of on going studies on the association between genetic and biochemical haemostatic variables and cardiovascular disease has been started and the forms to apply are now available at the above website. The forms can be completed and sent to Dr. Iacoviello directly through the website.

The Subcommittee's website is also linked to the CANVAS website  ( http://genecanvas.idf.inserm.fr/ ) which was presented by Dr. Cambien at last year's SSC meeting in Paris. CANVAS has the objective to facilitate the study of the impact of candidate gene polymorphisms on common cardiovascular disorders by accelerating the communication of information on candidate genes and DNA resources. It is also a single nucleotide polymorphism (SNP) resource that may be useful to those exploring the genomic regions where the candidate genes are located. In the next 5 years, 300 candidates genes are foreseen to be included in the catalogue. At the moment information on 566 polymorphisms, 114 genes and 29 studies are available.

Dr. Iacoviello also reported on the ETRO Working Party meeting on Population Genetics of Haemostatic Risk Factors for Arterial Vascular Disease that was held in Rome, on October 26-27, 2001.  The aim of this Working Party is to bring together European investigators in genetics, environment and thrombosis with different expertise (molecular and cell biology, epidemiology, biochemistry) to evaluate the contribution of genetic and environmental components to the risk of arterial thrombosis, with a special emphasis on haemostatic factors. The hot topics of this year's meeting were pharmacogenetics and innovative epidemiological approaches, with particular attention to large cohort studies.

Finally it has been proposed to write guidelines for population association studies in genetics. Potential topics are: selection of controls, sample size, multiple comparisons, statistical analysis (permutation test), admixture bias, gene-environment interactions, selection of polymorphisms.

Are D-Dimer levels a suitable marker of cardiovascular disease? Methodological and epidemiological aspects. Dr. Gordon Lowe presented an overview of fibrin D-dimer in prediction of cardiovascular disease. While there are a large number of available commercial assays which give very different values for plasma D-dimer, its association with cardiovascular risk appeared consistent regardless of assay type. (Standardization of D-dimer assay is being performed by the Subcommittee on Fibrinolysis).  Variation on sample storage, and on repeated measurement after 4 years, appeared minimal. Epidemiological association of D-dimer included age, female sex, country, oral estrogen use, obesity, lack of leisure activity, varicose vein, and prevalent arterial disease or venous thromboembolism. A meta-analysis reported the relative risk of coronary heart disease as 1.7 (top third of D-dimer compared to bottom third). There were fewer studies of the predictive value of D-dimer for stroke or venous thromboembolism, but these showed consistent association with risk. These associations were not explained by acute phase reaction (e.g. C-reactive protein). D-dimer levels are normalised by oral anticoagulant therapy and merit evaluation in selection of patients (and in monitoring) of such therapy.

Homocysteine Evaluation: Problems and Predictive Value. Dr. Marco Cattaneo reported on the association of homocysteine levels and risk of cardiovascular disease (CVD). While the association between homocysteine and venous thromboembolism is well established, only retrospective case-control studies or prospective studies in patients who already had an ischaemic event demonstrated a consistent association between homocysteine and CVD. In contrast, prospective studies in healthy subjects gave negative results. As a consequence a huge debate is going on whether high homocysteine levels are cause or consequence of atherosclerosis. A study in patients with stroke demonstrated that, following the acute phase, there is a decrease in homocysteine levels, concomitant with the increase in C-RP levels; at longer term, homocysteine levels got back to normal levels, never increased.  Although a final conclusion  on this debate can only be derived from interventional studies, lowering homocysteine levels (whose results are expected in the next few years), there are still some arguments for a role of homocysteine in the risk of cardiovascular disease. Probably to express its risk potential, homocysteine requires to synergise with other risk factors. Moreover, in prospective studies, the power of the association between homocysteine and CVD decreases with the length of follow-up, suggesting problems related to a longer  sample storage. Homocysteine levels can be determined by genetic (MTHFR polymorphism) and environmental factors (vitamins). While there is no consistent evidence of an association between the MTHRF polymorphism and CVD risk, there is evidence for an independent association with vitamin B6.

Dr. Cattaneo also reported on the standardization of pre-analytical conditions for homocysteine evaluation. Homocysteine levels are stable over 6 hours in samples collected in both ACD and EDTA when stored on ice. After storage at room temperature, homocysteine levels increase about 13% if collected in EDTA and only 3 %, if collected in ACD. Therefore, when room temperature storage cannot be avoided, ACD must be used as an anticoagulant.

The ARIC Project: Haemostatic Variables, their genetic control and Prediction of ischaemic arterial disease. Dr. Kenneth Wu presented an overview of the results of the ARIC study, a prospective investigation of a USA cohort of 15,792 healthy subjects aged 45-64 years who were followed for 10 years. A nested case-control study vas performed by comparing 365 incident cases and 734 non-cases. Fibrinogen, factor VII, factor VIII, von Willebrand factor, protein C, APC resistance, antithrombin III, t-PA, PAI-1, plasminogen, activated factor VII, and XII, Protrhombin fragment F1+2, D-dimer, beta-thromboglobulin and soluble thrombomodulin levels were evaluated. Fibrinogen, plasminogen and D-dimer levels were significantly and independently associated with the risk of ischaemic events. Unexpectedly, sTM showed a negative association with the risk of CVD. No associations were found for the other haemostatic factors evaluated. Combined analysis of pro- and antithrombotic factors can provide a more precise estimation of the association: indeed, increased levels of ICAM-1, fibrinogen, F VIII, and vWF interact with low levels of sTM in increasing the risk of CVD, while in the presence of high levels of sTM they did not affect the risk.

Polymorphisms in several haemostasis factor genes (fibrinogen, F VII, F II, F V, F XIII, TM, platelet glycoprotein IIIa and Ib) were also evaluated in a larger population of 800 cases and 900 controls. Only TM and GP Ib polymorphisms were associated with the risk of CVD; however, the association between TM polymorphisms and CVD risk was present in African American subjects but not in Caucasians. This suggests that different ethnic populations should always be analysed separately.

Ethnic differences in genotype distribution and risk of cardiovascular disease: the case of the Japanese
Dr. M. Murata further discussed ethnic differences in genotype distribution by presenting the case of the Japanese population. In Japan the incidence of venous thrombosis after surgery and myocardial infarction is lower in both males and females, despite the high prevalence of smoking. Some polymorphisms related to such disease (F V Leiden, prothrombin, F XIII, GP Iib/IIIa) are absent in the Japanese population, while many others have a different frequency as compared to the Caucasian population. Some of them are increased, while some others are decreased, without any apparent relation with the potential risk effect. While the reduced incidence of venous thrombosis can at least in part be related to the absence of related polymorphisms, the difference in AMI is related more to different environmental habits, especially diet and to a different pathogenesis.  Indeed, 52 % of AMI in Japan can be attributed to coronary spasm, compared to 11% in Europe. Independent predictors of spasm in Japan are smoking and NO synthase polymorphisms.

Dr Murata also showed some results concerning PAF acetylhydrolase gene in a Japanese population. This enzyme inactivates PAF. A F/V polymorphism is associated with plasma activity of the enzyme, showing the VV genotype higher activity, FF no activity and VF intermediate activity. Carriers of VV genotype also showed reduced platelet activation after PAF, but not after ADP or collagen stimulation. The F genotype was associated with an increased risk of stroke, particularly at younger age.

Activation markers of coagulation and fibrinolysis in twins: heritability of the prethrombotic state. Dr. Robert Ariens reported on the heritability of the prothrombotic state. Heritability is the proportion of variance in a phenotype due to additive genes. Twin studies are a useful tool to evaluate the heritability of a phenotype by comparing monozygotic twins (100% genetics) with dizygotic twins (50% genetics). The levels of F VII, FXII, F XIII, Fibrinogen, vWF and PAI-1 showed a high inheritability. This result was consistent with those obtained in family studies. However, new polymorphisms in genes related to these phenotypes only account for a small part of their heritability (e.g., PAI-1 4G/5G polymorphism accounts for 2% of PAI-1 level heritability, Fibrinogen beta chain polymorphism for 1% of fibrinogen heritability and F VII Arg353Gly polymorphism for 16% of F VII levels heritablity). In a sample of 115 twin pairs (59 MZ and 56 DZ), it has also been demonstrated that the levels of coagulation activation markers are also inherited. In particular, prothrombin fragment F1+2, TAT and D-dimer levels showed a degree of heritability of respectively 45, 40 and 65%. Also fibrin structure shows a 39% heritability, although the effect of environmental factors is also important. Among others, glycosilation can be an important determinant of fibrin structure in diabetic patients.

Genetic regulation of Factor VII and Factor XII: insights from the GAIT Study. Dr. John. Blangero provided an update of the results of the GAIT study, a project aimed at identification of QTL in large pedigrees of   Spanish families. He suggested that, before studying the genetics of discrete traits (diseases) that require very powerful methods, a useful approach is to identify genetic loci that regulate continuous traits (factor levels), which are called QTL (quantitative trait loci). The study of large pedigrees is one of the more powerful tools for this purpose and allows both the localization and the identification of a QTL.

The GAIT study includes 398 subjects from 12 thrombophilic and 9 randomly selected families, with a total of 57 cases of thrombosis. F VII and F XII liability has been studied. In a first phase, QTL F XII and F VII have been identified by using a wide genome scan with 363 genetic markers. High lod score (LD) has been found in chromosomes 5, 10 and 2 for F XII levels.  Chromosome 10 contains F XII gene, however, in chromosome 5 there should be another gene strongly influencing F XII gene.  Since a 46C/T polymorphism in F XII gene has been associated with the levels of F XII,  if whether such a polymorphism could be responsible for the large linkage signal found has been tested. After conditional analysis, the LD for chromosome 10, although decreased, remained still significant, suggesting that other polymorphisms can be relevant.

For F VII levels a significant QTL has been found in chromosome 13, that actually contains F VII gene. F VII gene was resequenced: all its polymorphisms (49) were identified with an allele frequency between 0.01 and 0.37. Posterior probability of functionality was analysed and 7 polymorphisms were identified as functional.  Conditioned analysis including these polymorphisms showed that the LS for F VII levels was reduced to a non-significant level, suggesting that F VII gene was completely dissected.

Finally, Dr. Blangero commented on the limitation of the case-control approach for genetic studies, based on linkage disequilibrium (LD) mapping. Indeed, LD is too unpredictable and, therefore, negative studies give no information.

Genetic polymorphisms of haemostatic factors: a word of caution. Dr. Pier Mannuccio Mannucci discussed the need to use large samples in case-control studies to have enough power to detect association for polymorphisms in haemostatic genes. He presented the results of a case-control study performed in collaboration with the ANMCO (Italian Association of Hospital Cardiologist), including more than 1000 patients with AMI at young age (under 45 years) and a corresponding number of healthy controls matched for age and sex. Cases differed from controls for all common environmental risk factors, such as smoking, hypertension, diabetes, dyslipidemia, etc.  Cases also had a higher prevalence of family history of AMI in respect to controls; however, this difference was not explained by any of the polymorphisms of haemostatic genes studied. Indeed, no associations were found between F VII, FV, F II, F XIII, GP IIIa, etc., polymorphisms and the risk of AMI at young age. The same results were obtained when only females were taken into consideration.

Conclusion:
The subcommittee meeting highlighted the need to continue to study the role of haemostatic variables in the prediction of cardiovascular risk. Activation markers such as D-dimer levels could be particularly relevant, and their evaluation, also in relation to other well established risk factors, should be introduced in future epidemiological studies. On the other hand, new haemostatic markers are emerging as possibly predictive, such as plasminogen and soluble thrombomodulin and need to be tested in future studies.

A possible powerful approach emerges for genetic studies:
  1. QTL identification in family studies
  2. identification of all possible polymorphisms of the genes related to the QTL
  3. test for their functionality
  4. dissect residual inference
  5. assess the absolute or relative risk in population studies.
  6. design the population study with a correct calculation of the sample size, selection of cases and controls and respecting the homogeneity of populations.