Oral anticoagulant drugs have improved the prognosis of patients with thromboembolic disease. However, optimal oral anticoagulation control is usually hampered by significant interindividual variability coupled with a narrow therapeutic window. In the elderly, this variability is enhanced by concurrent medications, nutritional status and physical condition. Age itself correlates with increased severity of complications, and adverse events due to anticoagulant drug are thought to be responsible for up to 5000 deaths a year in France.

Although half the patients receiving anticoagulant treatment are over 80, there has been only one study targeting this population so far. We therefore propose to study the pharmacokinetic (PK) and pharmacodynamic (PD) components of the response to fluindione, the main oral anticoagulant used in France, in patients over 80.

150 patients beginning fluindione treatment (or resuming after 2 weeks rest) will be recruited in the following departments: Geriatric, Cardiology, Nephrology, Cardiac Surgery, Internal Medicine (CHU Bichat-Claude Bernard, Paris) and Metabolic Diseases and Internal Medicine (CHU d'Angers). Blood sampling will take place before the beginning of the study (J0) to measure baseline INR and coagulation factors II and VII. Fluindione concentrations will also be measured in non-naive patients. INR, coagulation factors and fluindione will be measured at J2, J4, J6, J8, and twice weekly until they leave the hospital or up to a month. Many covariates will be recorded: age, gender, concurrent medications, biochemical analyses, functional and nutritional status. We will also investigate genetic factors by collecting DNA to genotype polymorphisms related to the target of the drug. Recent work has shown that both the response to common anticoagulant drugs and their metabolism was influenced by genetic polymorphisms, and there is now convincing evidence that drug targets are controlled by genetic polymorphisms which can play a major role in the variability of the response.

Throughout the study, the physicians remain free to adapt drug regimen and prescribe additional INR measurements.

Data analysis will be performed in INSERM Unit 738 (CHU Bichat-Claude Bernard) using nonlinear mixed-effect models, statistical techniques allowing the analysis of sparse data while quantifying the sources of variability.

We expect to gain a better understanding of the role of age, nutritional status, genetic factors and drug interactions in the variability of the response to fluindione. We will also assess whether measuring the activity of coagulation factors helps to anticipate dangerous increases in INR. These goals are vital to provide better care of the elderly and minimise costs arising from the frequency of severe side-effects.

Future perspectives include the development of a software and recommendations to help adapt anticoagulant treatment in the elderly, taking into account their condition.