Drug induced anaphylaxis is frequently attributed to the use of muscle relaxants during anaesthesia. Recently The Norwegian Medicines Agency recommended that rocuronium bromide (Esmeron) be withdrawn from routine practice due to frequent reports of anaphylaxis. Over a period of two and a half years approximately 150 000 patients received rocuronium as part of their anaesthesia. In this period the Norwegian drug authorities received 29 reports of anaphylaxis or anaphylactoid reactions in patients treated with rocuronium. This is in stark contrast to the situation in other Nordic countries where a total of only seven cases of anaphylaxis in approximately 800 000 patients treated with rocuronium had been recorded by December 2000. This situation highlights the many potential problems of the surveillance of adverse drug reactions: reporting bias may lead to an over-estimate of the risk of one drug compared to another, and the possibility of under-reporting of adverse events (due to a weak reporting culture) further limit the validity of such comparisons. The surveillance of adverse drug reactions also represents a statistical challenge. While adverse event reports may help us to estimate the anaphylaxis rate we need to appreciate the uncertainty of such estimates. Adverse reactions are rare, random, and mostly independent events, resulting from the successive exposure of patients to a low risk intervention. The frequency distribution of adverse events will therefore conform to that of a Poisson process. The resulting Poisson distribution may inform us about the variability of adverse event data. An understanding of these methodological problems and statistical challenges will allow anaesthesiologists to make informed decisions concerning the use of muscle relaxants and other drugs associated with severe adverse reactions.