MALD model helps predict death versus recovery for APAP overdose

March 16, 2012

A mathematical model that uses laboratory values commonly available on admission can help physicians estimate acetaminophen (APAP: N-acetyl-para-aminophenol) overdose amount, time elapsed since overdose, and potential outcome, according to a study published online February 13 in Hepatology.

A mathematical model that uses laboratory values commonly available on admission can help physicians estimate acetaminophen (APAP: N-acetyl-para-aminophenol) overdose amount, time elapsed since overdose, and potential outcome, according to a study published online February 13 in Hepatology.

Because timely administration of N-acetylcysteine (NAC) to prevent life-threatening liver injury is critical, and patients who overdose on acetaminophen often arrive at the hospital confused or comatose, Christopher H. Remien, University of Utah, Salt Lake City, and colleagues developed the Model for Acetaminophen-induced Liver Damage (MALD).

MALD uses a patient's aspartate aminotransferase, alanine aminotransferase, and international normalized ratio measurements on admission.

“Based on the extent of estimated liver injury, the model predicts death for patients who took over 20g of APAP without NAC administration within the first 24 hours,” the authors said.

Between January 1, 2006, and December 31, 2009, the investigators retrospectively tested the model on 53 patients discharged from the University of Utah.

They found that by using only these initial measurements, the model accurately predicted death versus recovery with 75% sensitivity, 95% specificity. Because the model does not describe kidney damage, the researchers added serum creatinine level in excess of 3.4 mg/dL as an additional criteria. When this value was added, sensitivity increased to 100%.

“Using only data available on admission, the model results fit the post-treatment time series of the markers of liver damage for the majority of individual patients,” the authors wrote. They conclude that the model compares favorably to statistical methods and should be validated in multicentric retrospective and prospective evaluation.