VALUTAZIONE DELLA SICUREZZA DELLA TERAPIA CON CUMARINICI IN PAZIENTI CON CIRROSI EPATICA E TROMBOSI DEL CIRCOLO PORTALE.

Background and Aims: Anticoagulation with vitamin K antagonists (VKAs) is an effective and relatively safe therapy for patients with portal vein thrombosis (PVT). However, the haemorrhagic risk of VKAs in relation with the presence of cirrhosis, has poorly been investigated Methods: We compared the VKAs-related bleeding risk in cirrhotic patients with de novo PVT (PVT-cohort, n=62) vs non-cirrhotic patients with a thromboembolic event (TE-cohort, n=160). Any bleeding during four years of follow-up or up to withdrawal of anticoagulation therapy, was recorded. The quality of anticoagulation control was measured by the time in therapeutic range (TTR) of the INR. Bleeding risk due to portal hypertension (PHT) in the PVT-cohort was compared with an independent series of cirrhotics with PHT unexposed to VKAs during follow-up (CH-cohort, n=53). Major bleeding episodes under anticoagulation were intracranial or retroperitoneal events, fatal bleeding events, need of hospitalization or transfusion, otherwise they were considered minor bleedings. All patients with cirrhosis were under prophylaxis for PHT-related bleeding according to current guidelines. Results: TE-cohort and PVT-cohort were comparable for age, sex. The mean of TTR was 67.7+/-20.9% for the former, 70.5+/-19.1% for the latter (p=0.379) but treatment with VKAs was longer for the TE-cohort (31.1+/-16.9 vs 23.0+/-16.2 months, p=0.001). Overall, 41 patients under anticoagulation experienced a bleeding episode (14 major/27 minor). The actuarial probability of major/minor bleedings was higher in PVT-cohort (23%major/30%) than in the TE-cohort (6%/20%) (p<0.001). However, the risk of upper-gastro-intestinal bleeding in PVT-cohort (15%) was the same as in the CH-cohort (13%) also adjusting for potential confounders, confirming the lack of impact of VKAs on the risk of bleeding due to PHT. Finally, the exclusion of the upper-gastrointestinal bleeding in the PVT-cohort led to a significant reduction of major bleedings accountable for VKAs, leaving a significant residual risk only for minor bleeding episodes (p<0.05). Conclusions: VKAs expose patients with cirrhosis and PVT to an additional risk of minor bleedings. This should be taken into account in future clinical studies to ameliorate the benefit/risk ratio of anticoagulation in this clinical setting.