A French trial randomly assigned 197 patients with cirrhosis who were undergoing transjugular intrahepatic portosystemic shunt (TIPS) for intractable ascites or prevention of variceal rebleeding to receive rifaximin, 600 mg twice daily, or placebo beginning 14 days before the procedure and continuing 168 days afterward. Overall, 34% of patients who received rifaximin before and after TIPS placement had hepatic encephalopathy (HE) versus 53% of those who received placebo, with no differences in adverse events or transplant-free survival.
The study was published online on Feb. 2 by Annals of Internal Medicine and was summarized in the Feb. 26 ACP Gastroenterology Monthly. The following commentary by Ronald L. Koretz, MD, FACP, appeared in the ACP Journal Club section of the July Annals of Internal Medicine.
HE is a predictable consequence of creating portosystemic shunts in patients with cirrhosis because more portal vein blood flow is diverted from the liver. As the presence of such shunts often disqualifies patients from entering clinical trials, limited data exist regarding how HE should be treated in this situation. An earlier randomized clinical trial assessing the use of rifaximin in patients with TIPS failed to find a benefit. Nevertheless, when HE does occur, the treatment is the same as that used when HE arises in patients without shunts, namely nonabsorbable antibiotics or disaccharides.
The trial by Bureau and colleagues was larger and longer than previous trials. Postulating that the benefit of rifaximin was due to inducing an alteration in the intestinal microbiome, patients began therapy 2 weeks before TIPS to create a more favorable intestinal environment. It is not clear whether this factor accounted for the benefit. Nonetheless, this is currently the best evidence to justify using rifaximin in HE associated with TIPS.
Although rifaximin is expensive, the reduced health care costs (e.g., less need for hospitalization and lower outpatient expenses) suggest that it is cost-effective. These estimates were based on the reduction in episodes of overt HE. Bureau and colleagues considered “isolated asterixis (asterixis as the only manifestation of HE)” to be overt HE. I was unable to find any more details about this condition in the online protocol or in the West Haven criteria. The cardinal feature of HE is cognitive dysfunction; asterixis is not specific for HE and, in the absence of problems in thinking and behavior, would not be a reason for hospitalization. If an important component of the benefit of rifaximin was solely to prevent asterixis in patients without concomitant cognitive dysfunction, my enthusiasm for this treatment would be dampened substantially.