Commentary: non‐haemodynamic effects of beta‐blockers in cirrhosis – more than meets the eye?

In addition to their benefit for prevention of variceal bleeding, recent evidence points to non-haemodynamic effects of nonselective beta-blockers (NSBBs) in decreasing intestinal permeability to microbes and thus reducing the risk of infection. However, clinical studies have shown mixed results. A meta-analysis including three RCTs showed a significant decrease in risk of SBP in patients on NSBBs; however, a prospective cohort study of patients with refractory ascites showed increased mortality in patients on NSBBs. Bajaj and colleagues address this controversy in a large retrospective cohort study of veterans with cirrhosis. Propensity score matching was used to compare patients on beta-blockers with controls. This powerful technique allows patients to be matched on a number of important variables to isolate the effect of NSBB therapy and in effect, model a randomised study. Despite an apparently lower rate of infections in compensated and decompensated cirrhotics using NSBBs, no difference was found once patients were matched on propensity score. The time to death or transplant was significantly shorter in patients on NSBB; however, propensity matching was not performed for this outcome. The major strengths of this study are the large number of patients and the use of propensity scores to ensure that comparisons were made between patients with similar characteristics. However, this methodology is not foolproof. It is unclear, for example, why patients with high propensity scores for NSBB use were not on NSBB therapy. This may be an important confounder. For example, patients in this group may have had poor follow-up or had EVL (endoscopic variceal ligation) instead, either of which may have affected the outcomes. In addition, the study only looked at in-hospital infections and could not account for patients who stopped NSBBs due to intolerance. Ultimately, this study suggests that NSBBs have minimal effect on the rate of infections and an unclear effect on mortality. An RCT would be required to answer the question definitively, but is unlikely to be deemed ethical due to the proven efficacy of NSBBs in variceal haemorrhage prophylaxis. Follow-up of previous randomised trials of NSBBs vs. EVL may be useful to assess the risk of infection with these therapies but would like be underpowered to definitively answer this question. Thus, although these data reassure and support the continued use of NSBB in patients with varices, some looming uncertainty remains.