There is a continuing need to identify measurable causes of the substantial variability of the individual response to oral anticoagulants, such as warfarin or acenocoumarol. Finding such causes is required to improve predictability and consequently reduce the high levels of morbidity and mortality associated to these treatments. Together with anthropometric parameters, the known polymorphisms of CYP2C9 and VKORC1 enzymes currently account only for about 50% of the dose variability [1–3].
The relation between measurable sources of variability and the variability of the patient response to treatment is explained by the complex but extensively explored warfarin pharmacokinetics/pharmacodynamics (PK/PD).
In short, Warfarin is absorbed from the gut completely, then it blocks irreversibly the Vitamin-K epoxyde reductase complex 1 (VKORC1), an enzyme found mostly in the liver. It is eliminated by other liver enzymes, from the cytochrome P-450 family, mostly by CYP450-2C9. The blocking of the VKORC1 results in a depletion of reduced Vitamin-K that is oxidised by the γ-glutamyl-carboxylase enzyme (GGCX) symultaneously with the γ-carboxylation of the glutamate residues of protein precursors of coagulation factors II,VII,IX and X. The γ-carboxylation is necessary for their biological activity.
The variability of this PK/PD process can be due to: genetic mutations of CYP2C9, VKORC1 and GGCX; mutations in the introns, exons or flanking sequences of these enzymes resulting in different phenotypical expression; variability in intake, transport and metabolism of vitamin-K; the size of the various organism compartments and in particular of the liver; the general synthesis capacity of the liver; simultaneous administration of drugs and foods that interact with various stages of the process, for example by inducing or repressing the expression of CYP450 enzymes.
The synthesis rate of coagulation factors, and their biological activity is also variable due to a number of mutations affecting both the structure and the expression rate of the precursors. The elimination rate is probably influenced by the general activation rate of the coagulation cascade in the bloodstream that also presents interindividual variability.
One potential cause of the response variability that has not been generally considered is that due to the genetic polymorphism and phenotypical expression of the proteins that intervene in the coagulation cascade. Their effect on the coagulation time, for example on the prothrombin time and its standardised equivalent--the International Normalised Ratio (INR), is low in the healthy population. Here we argue, based on a stochastic version of a well studied model of the extrinsic pathway of the coagulation cascade , that this effect should be low only in untreated subjects.
The canalization effect, [5, 6] is the relative robustness of key physiologic parameters, features and processes despite wide genetic and phenotypical variability of other factors, such as expression levels of proteins.
We believe this could be the case with the coagulation cascade: while there is substantial interindividual variability in the biological level of coagulation proteins, the cascade is organised such that the key blood coagulation time parameter is maintained within very narrow limits across the healthy population.
This would also explain why a very substantial reduction of the level of vitamin-K dependent factors (II,VII,IX,X) needs to be obtained in order to observe a clinically relevant prolongation of the prothrombin time.
However, our simulations show that once this reduction is achieved, the canalization phenomenon is also supressed and the interindividual variability of the expression of coagulation proteins in treated subjects becomes manifest and needs to be corrected by adjusting the Warfarin dose.