Criteria

King's College Criteria[64]

The most widely accepted prognostic tool for patients who present with ALF. They were developed through a retrospective analysis of 588 consecutive patients with ALF who were admitted to the King's College Hospital Liver Unit between 1973 and 1987.[64] Prognostic factors associated with mortality were identified and assessed for predictive value. Note that the international normalised ratio (INR) level used in this prognostic tool differs from the INR level used as a diagnostic feature of ALF.

Although fulfilment of these criteria has a high specificity for mortality, the sensitivity and negative predictive value remain low. Therefore, not fulfilling the criteria does not ensure survival.[81][82][83][84][85] The King’s College Criteria has a sensitivity of 68% to 69% and a specificity of 82% to 92%.[8] Although the King’s College Criteria have been validated in adult cohorts with ALF, data suggest they may not reliably predict outcomes in the paediatric population.[86]

Overall, these criteria are instrumental in selecting patients who have a high risk of mortality with ALF. However, they have limitations, and reliance upon prognostic scoring systems to determine candidacy for liver transplantation is not recommended by the American Association for the Study of Liver Diseases.[8]

ALF secondary to paracetamol overdose:

  • pH <7.30 or

  • INR >6.5 (PT >100 seconds) and serum creatinine >300 micromol/L (>3.4 mg/dL) in patients with grade 3 or 4 hepatic encephalopathy.

Non-paracetamol associated ALF:

  • INR >6.5 (PT >100 seconds), or

  • any 3 of the following: age <10 or >40 years; aetiology non-A, non-B hepatitis, or idiosyncratic drug reaction; duration of jaundice before hepatic encephalopathy >7 days; INR >3.5 (PT >50 seconds); serum bilirubin >300 micromol/L (>17.6 mg/dL).

Clichy criteria[62]

Based on a French prospective study of patients presenting with acute viral hepatitis, in which patients identified as having the lowest survival without liver transplantation included those with hepatic encephalopathy and low factor V levels.[62] These criteria predicted mortality with a positive predictive value of 82% and negative predictive value of 98% in this cohort. However, subsequent studies have reported much lower predictive values which were inferior to the King's College Criteria in other populations, including paracetamol and non-paracetamol ALF.[81][87]

Presence of hepatic encephalopathy and factor V level:

  • <20% of normal in patients aged <30 years, or

  • <30% of normal in patients aged >30 years.

Model for End-Stage Liver Disease (MELD)[88][89]

Adopted by the United Network for Organ Sharing and the Organ Procurement and Transplantation Network organisation, the MELD score is well established as a validated predictive model of short-term mortality in patients with cirrhosis and is currently utilised in the allocation of donor organs in patients awaiting liver transplantation in the US.[88][89] Several retrospective studies have reported the MELD score to have similar predictive value to the King's College Criteria for mortality associated with ALF.[90][91][92][93] Prospective data from the US Acute Liver Failure Study Group revealed that a MELD score ≥30 in patients with paracetamol overdose had a high negative predictive value of 82%, such that patients with MELD scores <30 had a high probability of survival. In non-paracetamol ALF, a MELD score ≥30 had a positive predictive value of 81%, yet these values were not more accurate than the King's College Criteria.[8][94] Based on findings from a large meta-analysis, the MELD score could have a role in predicting hospital mortality, particularly in non-paracetamol ALF.[95]

Acute Physiology and Chronic Health Evaluation (APACHE) II[96]

The APACHE II scoring system was developed to predict mortality in patients of all disease categories admitted to intensive care units. The score comprises 12 common physiological and laboratory parameters, adjusted for patient age and underlying chronic health problems.[96] One prospective study in patients with paracetamol overdose found that an APACHE II score >15 was associated with high mortality and provided similar predictive value to the King's College Criteria, while another study found a score of ≥20 to be more predictive of mortality and need for liver transplant.[25][97]

Acute Liver Failure Study Group (ALFSG) Index[98][99]

A prognostic index was developed utilising a cohort of 250 patients enrolled in the ALFSG, and then validated in a separate cohort of 250 patients. Variables at the time of initial presentation that were found to have a strong association with mortality or need for liver transplantation included advanced coma grade, bilirubin, INR, elevated phosphorus, and serum levels of M30 antigen, a marker of apoptotic hepatocyte cell death. This index was found to have an overall sensitivity of 85.6% and specificity of 64.7% with no significant difference in performance between paracetamol and non-paracetamol ALF. Although this predictive index was superior to the King’s College Criteria and the MELD score, assessment of M30 antigen requires an enzyme-linked immunosorbent assay (ELISA)-based test and may not be readily available at most centres.[98]

An additional model predictive of transplant-free survival was developed from the ALFSG database, involving 878 patients, and was then validated in a cohort of 885 patients. Variables predictive of transplant-free survival included the degree of hepatic encephalopathy, aetiology of ALF, requirement for vasopressors, bilirubin, and INR. Aetiologies of ALF considered to be favourable in this model included paracetamol (acetaminophen) overdose, pregnancy, ischaemia, or hepatitis A. The model performed with a sensitivity of 37% and specificity of 95% in determining an 80% transplant-free survival within the validation cohort.[99]

West Haven criteria for hepatic encephalopathy[49]

  • Grade 1: subtly impaired awareness, sleep alterations, shortened attention span, impaired addition or subtraction, heightened mood or anxiety, oriented in time and space.

  • Grade 2: lethargy or apathy, disorientation for time, obvious personality change, inappropriate behaviour, dyspraxia, asterixis.

  • Grade 3: somnolence to semi-stupor, responsive to vocal stimuli, marked confusion, gross disorientation (disoriented in time and space), bizarre behaviour. Physical findings may include hyper-reflexia, nystagmus, clonus, and rigidity.

  • Grade 4: coma.

Use of this content is subject to our disclaimer