Objective: To determine whether the Acute Physiology and Chronic Health Evaluation (APACHE) II system for the measurement of severity of illness is able to provide an accurate risk of hospital death in patients with acetaminophen-induced acute liver failure or identify those patients needing transfer for possible hepatic transplantation.
Design: Data for admission (first 24 hrs) APACHE II scores and King's criteria for urgent transplantation were collected prospectively to compare the APACHE II system and the King's criteria for the prediction of death or need for transplantation.
Setting: A nine-bed specialist liver failure unit (LFU).
Patients: One hundred two consecutive patients admitted to the LFU with acetaminophen self-poisoning and a prolonged prothrombin time were studied.
Interventions: None.
Measurements and main results: An APACHE II score of > 15 points was associated with a high mortality (13/20 patients, five of whom survived following hepatic transplantation). There was no relation between APACHE II risk and outcome (mean APACHE II risk of death 0.8%, actual hospital mortality 16%). An APACHE II score of > 15 had a similar power of prediction of death as the King's criteria (sensitivity 82% and 65%, respectively; specificity 98% and 99%, respectively), when considering those patients who were transplanted as "deaths." An APACHE II score of > 15 was able to identify four more patients than the King's criteria on the first day of admission to the LFU.
Conclusions: The crude admission APACHE II score correlated well with mortality in patients with acetaminophen-induced acute liver failure. However, the calculated APACHE II risk of death, using the original drug overdose coefficient, was poorly calibrated. Since specialist liver scores are unfamiliar in the general intensive care setting, the use of an APACHE II score might earlier identify more patients at risk of needing a liver transplant, and hence, expedite appropriate transfer to a specialist liver unit.