BACKGROUND: The optimal route and duration of administration for N-acetyl-cysteine in the management of acetaminophen (paracetamol) poisoning are controversial. It has been stated on the basis of a selected post-hoc analysis that oral N-acetylcysteine is superior to intravenous N-acetylcysteine in presentations later than 15 hours. AIM OF STUDY: To investigate the efficacy of intravenous or oral N-acetylcysteine. PATIENTS AND METHODS: We analyzed a series of acetaminophen poisonings treated with a protocol including activated charcoal and intravenous N-acetylcysteine. The outcomes assessed included use of N-acetylcysteine, adverse effects of intravenous N-acetylcysteine, and the occurrence of hepatotoxicity (transaminase > 1000 U/L). We incorporated these results in a meta-analysis of previously reported series of acetaminophen poisonings to compare the outcomes from intravenous and oral N-acetylcysteine use. RESULTS: Of 981 patients admitted over 10 years, 4% (40) presented later than 24 hours and 10% (100) had concentrations of acetaminophen that indicated a probable or high risk of hepatotoxicity. The 30 patients who developed hepatotoxicity presented later, took larger amounts, had higher concentrations, and received N-acetylcysteine later than those who did not. No patients received a liver transplant but 2 patients died (one after referral to a transplant unit and one just before). Adverse reactions to intravenous N-acetylcysteine occurred in 6% (12/205) of patients but none prevented completion of the treatment. In the meta-analysis, those with probable or high risk concentrations had similar outcomes with intravenous (pooled n = 341) and oral N-acetylcysteine (pooled n = 1462) administration. Rates of hepatotoxicity for those treated within 10 hours (3 and 6%), late (10-24 hours: 30 and 26%), and overall (0-24 hours: 16 and 19%) were all similar. The proportion of patients classified as presenting later than 10 hours is much greater in the oral N-acetylcysteine studies (64%) than in many of the intravenous N-acetylcysteine studies (38%, 44%, and 63%). CONCLUSIONS: The differences claimed between oral and intravenous N-acetylcysteine regimes are probably artifactual and relate to inappropriate subgroup analysis. A shorter hospital stay, patient and doctor convenience, and the concerns over the reduction in bioavailability of oral N-acetylcysteine by charcoal and vomiting make intravenous N-acetylcysteine preferable for most patients with acetaminophen poisoning.