Can haematological tests predict cardiovascular risk? The 2005 Kettle Lecture

The risk of venous or arterial thrombosis is routinely assessed by clinical variables (risk factors) supplemented by measurement of blood lipids and glucose for arterial thrombotic events. Haematological tests that might play a role in risk prediction include haemostatic variables, haematocrit and inflammatory markers (erythrocyte sedimentation rate, plasma viscosity, white cell count). Recent epidemiological studies of these phenotypes and related genotypes are reviewed. For the risk prediction of first venous thrombosis, screening for thrombophilias in 'high-risk' situations does not appear clinically effective or cost-effective; with the possible exception of women considering oral hormone replacement therapy. General screening after a first venous event to predict recurrence (or risk in asymptomatic relatives) does not appear effective; with the possible exception of D-dimer, which requires further study. For risk prediction of first arterial thrombosis, screening adds little to prediction by current clinical risk scores. Screening of persons after a first arterial event, or with atrial fibrillation (e.g. with D-dimer for stroke prediction), requires further study. In conclusion, haematological tests have very limited roles in the prediction of cardiovascular risk, and should only be used according to evidence-based guidelines. The need for management studies is highlighted.