Purpose: In many Korean hospitals, serum acetaminophen concentrations in cases of overdose cannot be measured initially because of inadequate laboratory facilities. Under these circumstances, physicians base the administration of the antidote, N-acetylcysteine, on ingestion amounts as determined by initial history taking. We therefore examined the correlated between ingested amounts and serum acetaminophen concentrations. Methods: Medical records were reviewed retrospectively for patients who presented to the ED with acetaminophen overdose between January 2002 and March 2006. Fifty-nine patients were recruited and sixteen patients were excluded. The forty-three remaining patients were placed into either the high-risk or low-risk group based on their ingested amount (140 mg/kg), and were separately categorized into the toxic or non-toxic group based on their serum acetaminophen concentrations, according to the Rurnack-Matthew nomogram. Results: Ten patients (83.3%) among twelve in the high-risk group were found to have non-toxic serum concentrations, and just one patient (3.2%) among thirty-one in the low-risk group fell into the toxic group based on their serum concentrations. The sensitivity and specificity of risk stratification of the ingested amount as a predictor of intoxication requiring antidote therapy were 66.7% and 75.0%, respectively. Conclusion: This study suggests that the therapeutic decision for acetaminophen overdose should not be based solely on ingested amount only, but requires assessment of acetaminophen concentration.