We encountered cases in which the anticoagulant effects of warfarin (CYP2C9 substrate) were reversibly attenuated by the concomitant administration of rifampicin or bosentan, which are potent pregnane X receptor (PXR) ligands. The purpose of the present study is to report the previous case with rifampicin, and to evaluate the changes in the warfarin anticoagulant effects when withdrawing or switching bosentan treatment. The former is a case study of a 4-year-old girl undergoing warfarin treatment. The latter is a longitudinal study of 20 pediatric patients receiving stable warfarin treatment. The prothrombin time and international normalized ratio (PT-INR) values were extracted from the medical records and normalized by the daily-dose per body size as an index for the warfarin anticoagulant effects. Rifampicin treatment resulted in a 52.0% decrease in the anticoagulant index. On the other hand, 10 of 20 patients started bosentan and their anticoagulant index was reduced by a median of 2.00. Bosentan was withdrawn in 4 of 20 patients and their anticoagulant index increased by a median of 3.67. Six of 20 patients switched from bosentan to macitentan, which is considered not to activate PXR in clinical settings. However, switching from bosentan to macitentan resulted in a median of 2.25 reduction of the anticoagulant index rather than recovery of the response to warfarin. This study suggests not only the possibility of heterogeneity in the response to PXR activation and deactivation, but also the importance of long-term monitoring of drug–drug interactions when switching from bosentan to macitentan.