Discussion
This systematic review identified eight studies exploring the effect of statin initiation on the degree of anticoagulation in VKA users. Overall, initiation of simvastatin treatment in warfarin users lead to a slight increase in mean INR ranging from 0.15-0.65, which seemingly peaks about 4 weeks after initiation.[13, 14],[19] For the other studied statins (atorvastatin, pravastatin, rosuvastatin, and fluvastatin) the same tendency of slightly increased anticoagulation was also present, described as INR increases, reduced doses of VKA and increased risk of hospitalization with GI bleeding.
The primary strength of our study is the multiple broad systematic search strategies designed to include original data. All literature was assessed by two persons, ensuring validity of the literature selection.
While all the included studies concerned patients in VKA treatment initiating treatment with statins, the studies showed considerable heterogeneity, with regard to type of statin and statin dose. Furthermore, the included studies concerned patients in VKA treatment with warfarin, phenprocomarol and acenocoumarol treatment. Finally, the patient population in the included studies also differed. While most studies excluded patients taking other medication that could interact with VKA and coagulation,[13,14],[16-19] two studies did not report whether these patients were excluded or not.[15, 20] In the two included clinical trials, treatment with rosuvastatin and atorvastatin was initiated in patients without any clinical indication requiring cholesterol lowering therapy.[15, 16] It is possible that the pharmacological response to statins and the interaction with VKA in patients with normal cholesterol levels differs from patients with elevated cholesterol levels. Further, for some studies, the lack of information regarding the type of statin[20] and the possible use of statins in the control group[18] result in difficulties, when interpreting the findings of the studies. Among the six studies where outcome was change in INR/PTT,[20] only half provided any analytical characterization of the INR/PTT tests used in the studies.[15, 16, 19] This missing laboratory information might explain at least in part the differences in the observed increases in INR.[21] Due to the heterogeneity of the studies, a direct comparison by meta-analysis was not considered feasible.
The pharmacological mechanism of the potential interaction between VKA and statins is to our knowledge not fully described. In vitro data suggests that statins have the potential to enhance the pharmacological activity of warfarin by competitively inhibiting its CYP-dependent metabolism.[14] However, considering the substantial delay in the drug-drug interaction (~four weeks) and the unspecific effect of several statins on warfarin effectiveness, this drug-drug interaction might not be mediated through cytochrome P450 inhibition, which would be expected to lead to a faster onset of INR increase, as e.g. seen for azole antifungals.[22] An alternative explanation for the observed associations could be related to the effect of changes in cholesterol levels on warfarin metabolism. Cholesterol levels stabilize about four weeks after initiation of statins [23, 24] coinciding with the maximum impact of statin initiation on INR among patients treated with VKA. However, further data is required to support this hypothesis.