Elsevier

Journal of Hepatology

Volume 61, Issue 5, November 2014, Pages 1038-1047
Journal of Hepatology

Research Article
Development and validation of a prognostic score to predict mortality in patients with acute-on-chronic liver failure

https://doi.org/10.1016/j.jhep.2014.06.012Get rights and content

Background & Aims

Acute-on-chronic liver failure (ACLF) is a frequent syndrome (30% prevalence), characterized by acute decompensation of cirrhosis, organ failure(s) and high short-term mortality. This study develops and validates a specific prognostic score for ACLF patients.

Methods

Data from 1349 patients included in the CANONIC study were used. First, a simplified organ function scoring system (CLIF Consortium Organ Failure score, CLIF-C OFs) was developed to diagnose ACLF using data from all patients. Subsequently, in 275 patients with ACLF, CLIF-C OFs and two other independent predictors of mortality (age and white blood cell count) were combined to develop a specific prognostic score for ACLF (CLIF Consortium ACLF score [CLIF-C ACLFs]). A concordance index (C-index) was used to compare the discrimination abilities of CLIF-C ACLF, MELD, MELD-sodium (MELD-Na), and Child-Pugh (CPs) scores. The CLIF-C ACLFs was validated in an external cohort and assessed for sequential use.

Results

The CLIF-C ACLFs showed a significantly higher predictive accuracy than MELDs, MELD-Nas, and CPs, reducing (19–28%) the corresponding prediction error rates at all main time points after ACLF diagnosis (28, 90, 180, and 365 days) in both the CANONIC and the external validation cohort. CLIF-C ACLFs computed at 48 h, 3–7 days, and 8–15 days after ACLF diagnosis predicted the 28-day mortality significantly better than at diagnosis.

Conclusions

The CLIF-C ACLFs at ACLF diagnosis is superior to the MELDs and MELD-Nas in predicting mortality. The CLIF-C ACLFs is a clinically relevant, validated scoring system that can be used sequentially to stratify the risk of mortality in ACLF patients.

Introduction

Acute-on-chronic liver failure (ACLF) is a syndrome characterised by acute decompensation of cirrhosis, organ failure(s) and high short-term mortality [1], which was recently defined in the CANONIC study [2]. This was a large prospective observational investigation carried out in 29 European university hospitals. It included 1349 consecutive patients admitted with acute decompensation of cirrhosis (ascites, bacterial infection, gastro-intestinal haemorrhage, and hepatic encephalopathy) and followed-up for one year. The CANONIC study was organised in the setting of the European Association for the Study of the Liver – Chronic Liver Failure Consortium (EASL CLIF-C).

The results of the CANONIC study showed that ACLF occurs most frequently in relatively young individuals, affects approximately 30% of hospitalised cirrhotic patients, may develop in patients without previous decompensation, is associated with a 28-day mortality rate of 33% (51% at 90 days) and is distinct from ‘mere’ decompensation of cirrhosis. ACLF is the most frequent indication for admission to the intensive care unit (ICU) [3]. In the US, about 200,000 patients with cirrhosis are hospitalised each year of which about 26,000 patients require ICU care [1], [3], [4]. An average ICU admission costs about $116,200 and costs for the health care system are $3 billion for cirrhotic patients requiring intensive care [3].

The diagnostic criteria for ACLF in the CANONIC study were based on the Chronic Liver Failure-SOFA score (CLIF-SOFAs), an adaptation for cirrhotic patients of the sepsis organ failure assessment score (SOFAs) widely used in the ICU [4]. The CLIF-SOFAs, however, is complex (based on 6 subscores, each with a 5-point range, assessing liver, kidney, brain, coagulation, respiration, and circulation), is based on consensus and expert opinion rather than data, and does not significantly improve the prediction accuracy of the model for end-stage liver disease (MELD) and MELD-sodium (MELD-Na) scores [5], [6].

The current study was aimed to simplify the original CLIF-SOFAs and develop a new score for ACLF patients (CLIF Consortium ACLF score, CLIF-C ACLFs) with a higher prognostic accuracy than the CLIF-SOFA, MELD, MELD-Na, and the Child-Pugh (CP) scores [7] for patients with ACLF. The study therefore had four main objectives. First, to develop a simpler and validated organ failure score (CLIF Consortium Organ Function score, CLIF-C OFs) for the diagnosis and grading of ACLF. Second, to design a more accurate prognostic score for ACLF patients (CLIF-C ACLFs), using the CLIF-C OFs and other prognostic clinical and biochemical data. Third, to compare the prognostic accuracy of the CLIF-C ACLFs to that of MELDs, MELD-Nas, and CPs. Fourth, to validate the prognostic accuracy of the CLIF-C ACLFs in an external prospective cohort of consecutive patients hospitalized in a single ICU and assess the score for sequential use.

The CANONIC study database was used as a derivation set for several reasons: first, it includes a large series of patients with acute decompensated cirrhosis and also with ACLF. Second, CANONIC patients were closely and prospectively followed-up for up to 1-year. Third, the population with ACLF included patients developing the syndrome either at study inclusion or during the hospitalization period. Finally, as patients were recruited from 29 centres in Europe, the CANONIC data are representative of the European patient population.

Section snippets

Study populations

The study was performed in patients from three different populations. Both the derivation and the validation datasets came from studies approved by Ethical Review Boards of all study sites.

  • 1.

    The CLIF-C OFs was developed using the baseline (i.e., enrolment) data of the whole CANONIC study population, which included 1349 patients (out of 2149 screened) admitted to 29-European hospitals within a period of 6 months for the treatment of decompensated cirrhosis. These patients were prospectively

ACLF study populations

Table 1 shows the two ACLF study populations. ACLF patients in the external validation cohort were more frequently male and alcoholic, had more severe ACLF and higher rates of ascites, hepatic encephalopathy, respiratory and circulatory failure than ACLF patients in the CANONIC cohort. Consequently, the MELDs and MELD-Nas and the 28-day and 90-day mortality rates were higher in the validation cohort.

Development of the CLIF-C OFs

For each organ system, two new cut-points were chosen to distinguish three clinical severity

Discussion

The observation in the CANONIC study that ACLF occurs in 30% of hospitalised cirrhotic patients, is associated with a 28-day mortality rate of 32%, has well-defined diagnostic criteria, and is pathophysiologically related to systemic inflammation provides the rationale for new investigations on the mechanism and management of this syndrome [1], [2], [8], [15], [16]. Consequently, it is necessary to develop an accurate prognostic score for ACLF patients.

Any specific score for ACLF should have

Financial support

The CLIF Consortium is supported by an unrestricted grant from Grifols. Rajiv Jalan is supported by a comprehensive biomedical research centre, UK grant, Richard Moreau was supported by an INSERM-APHP fellowship and Jonel Trebicka by grants from the Deutsche Forschungsgemeinschaft (SFB TRR57 project 18).

The EASL-CLIF Consortium

The EASL-CLIF Consortium is a network of 63 European university hospitals, aimed at stimulating research on pathophysiology, diagnostic and treatment on Chronic Liver Failure. During the period 2009–2012 the EASL-CLIF Consortium had received unrestricted grants from Grifols and Gambro. Grifols has prolonged its unrestricted grant for an additional period of four years. There is no other support for the Consortium. The Fundació Clinic, a foundation ruled by the Hospital Clinic and University of

Authors’ contributions

R. Jalan, R. Moreau, M. Pavesi, A. Amoros, F. Saliba, M. Bernardi, and V. Arroyo participated in the writing group. R. Moreau, R. Jalan, P. Ginès, M. Pavesi, F. Durand, T. Gustot, J. Trebicka, and V. Arroyo designed the CANONIC study with input from all other authors. R. Moreau, R. Jalan, M. Pavesi, A. Amoros, F. Saliba, and V. Arroyo participated in data analysis and interpretation. F. Durand, P. Angeli, P. Caraceni, C. Hopf, C. Alessandria, E. Rodríguez, P. Solis-Muñoz, W. Laleman, J.

CANONIC study investigators of the EASL-CLIF Consortium (alphabetical order)

  • 1.

    Patricia Aguilar Melero, Hospital Universitario Reina Sofía. Centro de Investigación Biomédica en Red Enfermedades Hepáticas y Digestivas (CIBERehd). Instituto Maimónides de Investigación Biomédica de Córdoba (IMIBIC), Córdoba, Spain.

  • 2.

    Agustin Albillos, Servicio de Gastroenterología, Hospital Universitario Ramón y Cajal, Universidad de Alcalá, Centro de Investigación Biomédica en Red Enfermedades Hepáticas y Digestivas (CIBERehd), Madrid, Spain.

  • 3.

    Rafael Bañares, Hospital General Universitario

References (19)

There are more references available in the full text version of this article.

Cited by (734)

View all citing articles on Scopus

These authors contributed equally to this work.

The complete list of CANONIC study investigators is reported at the end of the manuscript.

View full text