A prospective study of 12-month survival of patients with cirrhosis admitted to intensive care.

BACKGROUND

The assessment on whether to escalate care in patients with cirrhosis to the intensive care unit (ICU) is complex, with little data on long-term outcomes. The primary aim of this study was to prospectively examine the ability of liver-specific and ICU prognostic scores to predict 12-month survival among patients with cirrhosis in one non-transplant health service. The secondary aim was to assess the prognostic scores in predicting shorter-term survival and the correlation between other demographic and biochemical factors.

METHODS

All patients with cirrhosis who had an unplanned admission to the ICU were included in the study prospectively. Patient demographics and biochemical markers were collected, liver-specific and ICU prognostic scores, as well as the Chronic Liver Failure Consortium acute liver failure and acute decompensation scores were calculated. The pre-acute deterioration Child Turcotte Pugh, Model for End-Stage Liver Disease (MELD) and MELD-Na scores from the 1-6 months prior to the admission were also obtained.

RESULTS

One hundred patients were included over 33 months, with a median follow-up of 25 months. Sixty percent of patients survived longer than 12 months. Sepsis was the most common cause of cacute on chronic liver failure; (ACLF), which was associated with a lower 30-day survival. Patients with non-variceal gastrointestinal bleeding were more likely to survive 12 months. In patients with ACLF (n = 57), the day 1 MELD score was the best at predicting long-term survival, whereas in patients without ACLF, none of the scores performed particularly well, with an area under the receiver operating characteristic curve below 0.6.

CONCLUSION

Patients with cirrhosis may have excellent long-term outcomes following ICU admissions. Distinguishing whether a patient is in ACLF or non-ACLF may aid with determining the appropriate prognostic score to apply to aid decision making on limitations of patient care.