The anticoagulant warfarin is one of the most widely prescribed drugs in modern medicine. To achieve successful anticoagulation and to avoid life-threatening bleeding, patients need their dosage to be individualized when they start the drug and then to be monitored regularly to ensure they are still receiving the correct dose. Our paper was the first to show that variant CYP2C9 alleles were more common in patients who needed a low dose of warfarin and suggested that it might be helpful to genotype patients for CYP2C9 prior to the start of drug treatment. Individualizing drug treatment on the basis of genotype for pharmacogenetic polymorphisms such as CYP2C9 is of considerable current interest, with the relationship between CYP2C9 genotype and warfarin dose requirement a frequently cited example.

Our findings are of direct relevance to the use of warfarin as an anticoagulant and there is currently ongoing research in several centers worldwide on how best to incorporate genotyping for CYP2C9 and possibly other genes in routine clinical practice.

My colleague Chris Day and I had funding from the European Union to study the significance of cytochrome P450 polymorphisms in relation to drug-induced hepatotoxicity of the nonsteroidal anti-inflammatory drugs (NSAIDs). Guru Aithal, first author on the Lancet paper, worked with us as a Research Fellow. CYP2C9 is the main NSAID hydroxylating enzyme and we set up genotyping assays for the most common variant alleles as well as assessing their functional significance in relation to NSAID metabolism. As sample collection for the main study was proceeding quite slowly, we decided to initiate a "side project" on the relationship between CYP2C9 genotype and warfarin dose requirement since warfarin was a well-established CYP2C9 substrate. We involved Patrick Kesteven, a consultant hematologist, and decided to study individuals who required unusually low doses of warfarin to see if they were more likely to have variant CYP2C9 alleles which were already known to be associated with slower than normal metabolism of warfarin. We found that 81% of patients who needed a warfarin dose of 1.5 mg per day or less were positive for one or two variant CYP2C9 alleles, compared with 40% of either healthy controls or patients on warfarin generally which was highly statistically significant.

Our findings, which have been confirmed by other researchers in a number of subsequent studies, suggest that knowledge of a patient's CYP2C9 genotype is helpful in predicting the dose of warfarin needed for successful anticoagulation or "thinning" of the blood in patients at risk of blood clots. It may be possible in the future for patients needing to be prescribed warfarin to have their genotype tested before starting treatment and for their initial drug dose to be set on the basis of their genotype. This should lead to a decreased risk of suffering problems with bleeding and, for certain patients, a shorter hospital stay.