Compared to warfarin, little is known about the effect of pharmacogenomics on the inter-individual variability of phenprocoumon therapy. In a retrospective cohort study, we investigated the impact of VKORC1 c.-1639G>A; CYP2C9*2 , CYP2C9*3 ; GGCX c.214+597G>A; CALU c.*4A>G; EPHX1 c.337T>C; F7 c.-402G>A, and F7 c.-401G>T on the initiation (n=54) and maintenance phases (n=91) of phenprocoumon therapy. We assessed the following outcome parameters: time to stable international normalised ratio (INR), time to first supra-therapeutic INR, time out of INR range, probability of over-anticoagulation, number of anticoagulation clinic visits. During the initiation phase, homozygotes for the VKORC1 c.-1639 A and G alleles achieved stable INRs later (p<0.001), spent more time at supra-therapeutic INRs (p<0.001), had increased risks of over-anticoagulation (odds ratio 19.83, p=0.003 and 4.45, p=0.045, respectively), and had higher frequencies of anticoagulation clinic visits (p<0.001) compared to GA carriers. CYP2C9*2, *3 carriers reached stable INRs faster (p=0.024) with fewer anticoagulation clinic visits (p=0.001) than wild-type carriers. EPHX1 c.337 C carrier spent significantly more time above range in the initiation phase (p=0.023). GGCX , CALU , and F7 gene variants did not affect outcome parameters of the initiation phase and none of the genotypes had an impact on maintenance phase parameters. Compared to the VKORC1 genotype, early INR values were less informative in the prediction of outcome parameters such as time to stable INR and time above the INR range. Our study is limited by the retrospective study design with no standardised protocol in a usual care setting. Therefore, our findings should be validated in a larger, controlled prospective study.