De Gottardi et al.1 have reported a safe use of direct oral anticoagulants (DOACs) in patients with splanchnic vein thrombosis with or without cirrhosis. A few cases of portal vein thrombosis (PVT) resolution in cirrhotic patients after treatment with DOACs have been reported to date, whereas their efficacy in the prevention of PVT has not been investigated yet. We have recently observed a case of PVT occurred in a 81-year- old woman with cryptogenic cirrhosis, Child-Pugh score B8, under treatment with the DOAC rivaroxaban 20 mg daily for chronic atrial fibrillation. Portal vein thrombosis was diagnosed during the periodic liver ultrasound study and involved the splenomesenteric confluence, the portal vein trunk, and right and left portal vein branches. The last ultrasound examination performed 3 months before showed a patent portal vein but a reduced mean flow velocity (9 cm/s). Genetic or acquired thrombophilic factors were excluded. Peak rivaroxaban plasma concentration at 2 h after the assumption was within therapeutic limits (Cmax: 241 ng/mL), ruling out drug malabsorption. Rivaroxaban treatment was stopped and the patient was switched to low molecular weight heparin (LMWH), but developed portal vein cavernoma. In our experience, rivaroxaban was not able to prevent the occurrence of PVT. Although in the study by De Gottardi et al. DOAC serum levels were not provided, a recent in vitro study has reported a decreased anticoagulant potency of rivaroxaban in plasma of cirrhotic patients compared to samples from control subjects, whereas a similar anticoagulant potency of dabigatran, heparin and LMWH in patients and controls was shown.2 Previous studies showed that in patients with moderate liver dysfunction rivaroxaban plasma level 12 hours after intake of a single dose of 10 mg was higher than in healthy subjects. In our case, rivaroxaban plasma level two hours after the assumption was comparable to that reported for healthy subjects and for cirrhotic patients with mild hepatic impairment.3 However, the reduced portal vein flow velocity might have further favoured the occurrence of PVT in our patient.4 In conclusion, although preliminary evidences such as the study of De Gottardi et al. support a safe use of DOACs in cirrhotic patients, pharmacokinetics and pharmacodynamics may be altered and should be further and extensively investigated. We agree with the VALDIG Investigators that large studies are essential to definitively assess the role of rivaroxaban and other DOACs in the treatment and prevention of PVT in these patients.