Coumarin derivatives, including warfarin, acenocoumarol and phenprocoumon, are the drugs of choice for long-term treatment and prevention of thromboembolic events. The management of oral anticoagulation is challenging because of a large variability in the dose-response relationship, which is in part caused by genetic polymorphisms. The narrow therapeutic range may result in bleeding complications or recurrent thrombosis, especially during the initial phase of treatment. The aim of this review is to systematically extract the published data reporting pharmacogenetic influences on oral anticoagulant therapy and to provide empirical doses for individual genotype combinations. To this end, we extracted all data from clinical studies of warfarin, phenprocoumon and acenocoumarol that reported genetic influences on either the dose demand or adverse drug effects, such as bleeding complications. Data were summarized for each substance, and the relative effect of each relevant gene was calculated across studies, assuming a linear gene-dose effect in Caucasians. Cytochrome P450 (CYP) 2C9, which is the main enzyme for rate-limiting metabolism of oral anticoagulants, had the largest impact on the dose demand. Compared with homozygous carriers of CYP2C9*1, patients homozygous for CYP2C9*3 were estimated to need 3.3-fold lower mean doses of warfarin to achieve the same international normalized ratio, with *2 carriers and heterozygous patients in between. Differences for acenocoumarol and phenprocoumon were 2.5-fold and 1.5-fold, respectively. Homozygosity of the vitamin K epoxide reductase complex subunit 1 (VKORC1) variant C1173T (*2) allele (VKORC1 is the molecular target of anticoagulant action) was related to 2.4-fold, 1.6-fold and 1.9-fold lower dose requirements compared with the wild-type for warfarin, acenocoumarol and phenprocoumon, respectively. Compared with CYP2C9 and VKORC1 homozygous wild-type individuals, patients with polymorphisms in these genes also more often experience severe overanticoagulation. An empirical dose table, which may be useful as a basis for dose individualization, is presented for the combined CYP2C9/VKORC1 genotypes. Genetic polymorphism in further enzymes and structures involved in the effect of anticoagulants such as gamma-glutamylcarboxylase, glutathione S-transferase A1, microsomal epoxide hydrolase and apolipoprotein E appear to be of negligible importance.Despite the clear effects of CYP2C9 and VKORC1 variants, these polymorphisms explain less than half of the interindividual variability in the dose response to oral anticoagulants. Thus, while individuals at the extremes of the dose requirements are likely to benefit, the overall clinical merits of a genotype-adapted anticoagulant treatment regimen in the entire patient populations remain to be determined in further prospective clinical studies.