Summary Background Rifaximin might decrease the risk of portal hypertension‐related complications by controlling small intestinal bacterial overgrowth. Aim To evaluate whether rifaximin was associated with the risk of death and cirrhotic complications. Methods We conducted a retrospective study that included 1042 patients experiencing hepatic encephalopathy (HE): 421 patients without hepatocellular carcinoma (HCC; the non‐HCC cohort) and 621 patients with HCC (the HCC cohort). The primary endpoint was overall survival and secondary endpoints were recurrence of HE and the development of spontaneous bacterial peritonitis (SBP), hepatorenal syndrome (HRS) and variceal bleeding. Results In the non‐HCC cohort, 145 patients received rifaximin plus lactulose (the rifaximin group) and 276 patients received lactulose alone (the control group). The multivariate analysis revealed that rifaximin was significantly associated with lower risk of death (adjusted hazard ratio aHR, 0.697; P = .024) and reduced the risk of recurrent HE (aHR, 0.452; P < .001), SBP (aHR, 0.210; P < .001) and variceal bleeding (aHR, 0.425; P = .011) but not HRS (aHR, 0.598; P = .08). In the HCC cohort, 173 patients received rifaximin plus lactulose and 448 patients received lactulose. Rifaximin was not associated with the risk of death (aHR, 1.177; P = .121). Rifaximin was associated with lower risk of SBP (aHR, 0.323; P < .001) but not with variceal bleeding (aHR, 0.660; P = .104) or recurrent HE (aHR, 0.689; P = .057). The risk of Clostridium difficile‐associated diarrhoea was not different between the groups (aHR, 0.028; P = .338). Conclusions In patients without HCC, rifaximin treatment was significantly associated with prolonged overall survival and reduced risks of spontaneous bacterial peritonitis, variceal bleeding and recurrent hepatic encephalopathy. Linked Content This article is linked to Juang, Kang et al, Wang et al, and Borentain et al papers. To view these papers visit https://doi.org/10.1111/apt.14325, https://doi.org/10.1111/apt.14343, https://doi.org/10.1111/apt.14353 and https://doi.org/10.1111/apt.14516.