Metamizole was withdrawn from the market in the United States and several European countries following reports of fatal agranulocytosis among users, but is still available in many countries in Europe, South America and Asia. Over the past several decades, a number of epidemiologic studies have been conducted to quantify the risk of agranulocytosis and other adverse effects associated with metamizole and other non-narcotic analgesics. The objective of this study was to perform a systematic review of the safety of metamizole.Epidemiologic studies published between 1 January 1980 and 15 December 2014 were identified through systematic searches of PubMed and Google Scholar; the reference sections of selected articles were also reviewed to identify potentially relevant studies. Studies included in this review focused on the safety of metamizole, that is on outcomes such as haematologic abnormalities, gastrointestinal bleeding, anaphylaxis and hepatotoxicity. Two study investigators independently reviewed the abstracts and articles to determine relevant studies according to prespecified criteria.A total of 22 articles met the criteria for evaluation. The majority of studies that evaluated agranulocytosis indicated an increased risk associated with metamizole, with relative risk (RR) estimates ranging from 1·5 (95% CI, 0·8-2·7) to 40·2 (95% CI, 14·7-113·3). Findings of three case-control studies do not suggest an association between metamizole and aplastic anaemia. Of the five case-control studies that evaluated the risk of upper gastrointestinal bleeding, four found a statistically significant increased risk associated with metamizole (RR estimates ranging from 1·4 to 2·7). There is insufficient evidence to determine whether metamizole increases the risk of other outcomes (e.g. hepatic effects, anaphylaxis, congenital anomalies). Few studies evaluated the effects of dose, route of administration or duration of therapy.Published studies reported differences in the magnitude of risk of adverse outcomes associated with metamizole use and often had small sample sizes and a number of other limitations that may have biased the results. Further research is needed to better quantify the potential risks associated with metamizole compared to other non-narcotic analgesics.