Reply to: Correspondence on “Clinical features and evolution of bacterial infection-related acute-on-chronic liver failure”

Firstly, the authors would like to thank Dr. Sundaram and Dr. Fischer et al. for their interest in our paper. We also want to thank Dr. Fischer for providing some local data on a subgroup of patients with cirrhosis who were admitted to the intensive care unit (ICU). Comparing their patients to the entire group of patients in the global study, Dr. Fischer’s patients were a lot sicker, by virtue of the fact that they required ICU care. Many of them had multiple complications of cirrhosis, multiple infection sites, many more nosocomial infections (which are usually associated with a worse outcome), had a higher qSOFA score or sepsis. Therefore, it is not surprising that many more of them had higher grades of acute-on-chronic liver failure (ACLF), associated with significantly higher 28-day mortality of 68% compared to 37% in the GLOBAL study. Therefore, Dr. Fischer and colleagues are entirely correct that depending on their standpoint, the results are totally different. Study results can also depend on local epidemiology and clinical practice, as indicated by a subgroup of patients from the Indian sub-continent who had a higher rate and severity of bacterial infection-triggered ACLF than in Europe and America. We apologize for not having included many centres from Eastern Europe. Dr. Fischer’s information has now provided a glimpse of what infection triggered ACLF is like amongst ICU patients. Hopefully sometime in the future, we will be able to see some data from non-ICU patients from Eastern Europe.

With respect to Dr. Sundaram’s request, it is indeed true that we did not have too many patients from North America, and therefore the patients with non-alcoholic steatohepatitis (NASH) are under-represented. We have now included 2 tables providing data on patients with NASH as the etiology of cirrhosis. Table 1 compares patients with NASH who had ACLF irrespective of whether it was evident on admission or acquired during hospitalization to those with NASH who did not have ACLF throughout their entire hospital stay, whereas Table 2 compares NASH non-ACLF patients to those who developed ACLF after hospital admission. Table 1 tells us that essentially those patients with NASH-related ACLF were similar to those with ACLF related to other aetiologies of cirrhosis. The same outcomes were observed irrespective of whether the patients already had ACLF on admission or developed ACLF after acquiring infection during hospitalization. As NASH becomes more of a global problem, we will probably see a higher prevalence of NASH-related cirrhosis together with their many co-morbid conditions, such as diabetes and chronic kidney failure, making these patients more prone to the development of organ failure, ACLF and failure to respond to treatment. Therefore, careful antibiotic stewardship is ever more important in our daily care of these patients.