Prevalence and outcome of cirrhosis patients admitted to UK intensive care: a comparison against dialysis-dependent chronic renal failure patients: Reply to comments by Mason and Yeoman

Abstract

Dear Editor, We thank Drs Mason and Yeoman for their comments on our paper [1] as these raise a number of very interesting points. However, we would challenge the statement that these points mitigate against our conclusions. As mentioned in the previous reply to Drs Berry and Thompson [2], we do not wish to be seen as champions of nihilism for patients with cirrhosis; rather, our pessimistic conclusions simply reflect the data presented. Indeed, despite improvements in many aspects of critical care over the last 20 years, ICU mortality rates associated with cirrhosis remain very high. We felt that it was time for a change in the message from ‘‘we must try harder’’ to something bolder. It is very hard to advocate continuing ICU care when mortality rates exceed 85–90 %, especially as the vast majority of patients will not be eligible for transplantation. We perhaps should have added a section about optimising pre-ICU care and feel the authors’ critical care bundle suggestion has considerable merit, although whether this would have a benefit on mortality rates remains of course uncertain. However, the suggestion that cirrhosis patients be admitted earlier to the ICU is simplistic and naive given that using Hospital Episode Statistic data we identified [100,000 patients admitted to UK hospitals 2010–2011 with complications of cirrhosis (http://www. hesonline.nhs.uk). As an alternative, we would suggest borrowing another approach from sepsis studies—that of early ‘‘goal-directed therapy’’ [3], either in the emergency room or on the acute admissions ward. Indeed, a recent study has demonstrated that delayed appropriate initial empiric antimicrobial therapy is associated with increased mortality in cirrhosis [4]. As we acknowledge in our paper, the comparison with renal replacement therapy is imperfect, and we again refer to this in our reply to Drs Berry and Thompson [2], However, this hardly mitigates against our conclusion that cirrhosis patients with organ failure have an extremely poor prognosis. Rather than labour this point, we would refer these authors to the very similar overall mortality (59.2 %) seen in patients with haematological malignancies admitted to ICU [5], as mentioned in our discussion; this is a group we hope that all will agree can be regarded as having a poor prognosis. We cannot agree that using the Intensive Care National Audit & Research Centre (ICNARC) database to examine the largest cohort of ICU admissions with liver cirrhosis reported to date can be regarded as a study weakness. We acknowledge in our discussion that neither the Acute Physiology and Chronic Health Evaluation II (APACHE II) nor ICNARC score contains information directly relating to liver failure and conclude that neither score can guide decision-making; however, the area under the receiver operating characteristic curve (AUROC) of 0.8 that is seen with the ICNARC score is comparable to that obtained using Sequential Organ Failure Assessment (SOFA) and Organ System Failure (OSF) models in other studies. Equally, the fact that these scores under-predicted mortality had not been reported before, and we thought it worthy of mention. Again, we do acknowledge in the paper that mortality associated with variceal bleeding has improved over time; as such, a breakdown of mortality associated with differing reasons for admission may have proved interesting. However, the key determinants of ICU mortality in patients with cirrhosis are sepsis and organ failure [6] and, therefore, the aim of the study was to examine these in detail.