Use of metamizole in children and the risk of agranulocytosis

Is the benefit worth the risk?

This Invited Commentary accompanies the following original article:

Fieler M, Eich C, Becke K, et al. Metamizole for postoperative pain therapy in 1177 children. A prospective, multicentre, observational, postauthorisation safety study. Eur J Anaesthesiol 2015; 32:839–843.

In the current issue of the European Journal of Anaesthesiology, Fieler et al.¹ present a multicentre observational study that evaluates the use of metamizole in children up to 6 years of age undergoing surgery. The study paid particular attention to safety issues. One thousand one hundred and forty-five young patients were included (age 35.8 ± 18.1 months) and all received a single dose of metamizole (17.3 ± 2.9 mg kg⁻¹) for postoperative pain. Pruritus, swelling and exanthema were observed in a single patient only, reinforcing in the authors’ minds that the probability of serious adverse drug reactions (ADRs) is lower than 0.3% and that metamizole can be used safely in children under the age of 6 years.

We would like to comment on this assumption. Metamizole, also known as dipyrone, is a pyrazolone NSAID, known for its propensity to induce agranulocytosis. The first description of this life-threatening ADR of metamizole dates back to 1936.² Following numerous reports of agranulocytosis, some of which were fatal, metamizole was withdrawn from the market in the USA and the UK. In parallel, the use of metamizole in Europe decreased in the years that followed, but it is once again gaining popularity. The reason for this is the perception that the incidence of agranulocytosis with metamizole is not high enough to justify a ban on its use. This view is challenged by several studies that confirm a significant risk of agranulocytosis associated with metamizole. The Berlin Case-Control Surveillance Study (FAKOS), which evaluated a 10-year surveillance programme of 51 forms of nonchemotherapeutic drug-induced agranulocytosis, showed that the spectrum of drugs causing agranulocytosis was similar over the years, and that metamizole was still the leading cause in adults.^3,43,4 The reported incidences vary widely. A Swedish study found that about one prescription of metamizole in 1439 results in agranulocytosis,⁵ whilst another international study found a much lower risk of about 1.1 cases per 1 million patients.⁶ Calculations based on available data have suggested that metamizole is associated with at least 7000 cases per year of agranulocytosis worldwide.⁷

The mechanism of metamizole-induced agranulocytosis is thought to be idiosyncratic via immune-mediated mechanisms.⁸ Its latency, after introduction of metamizole, can range from 3 days to several months. A recent study⁹ of 858 reports of metamizole-associated haematological ADRs, wherein latency was available, found that half the cases appeared during the first week of treatment.

The WHO global database, named VigiBase, collects all individual case safety reports (ICSRs) sent by the national pharmacovigilance centres of currently 120 countries. In this database with more than 10 000 000 ICSRs, agranulocytosis with metamizole represents 4.6% of all agranulocytosis cases (1078 cases with metamizole and a total of 23 286 cases for all drugs on 15 February 2015). For comparison, by 15 February 2015, there were only 183 cases of agranulocytosis reported for diclofenac, another injectable NSAID, even though the total number of reports was four times that of metamizole, 65 789 for diclofenac and 14 080 for metamizole. Any conclusion should be approached with caution because there are other explanations for this reported increase in agranulocytosis with metamizole. These include the extent of use of the drug, the degree of ascertainment and the nature of the reactions. We should also note that we have no information on the number of patients exposed to any specific drug.¹⁰ But because this increase was not observed with diclofenac, it is probable that it is due to the drug and not secondary to the method of ascertainment.

Another point from the study of Fieler et al.¹ that we would like to stress is the absence of systematic haematological evaluation. The only appraisal of agranulocytosis in this study was through its versatile clinical expression; agranulocytosis was defined as unlikely in patients without persistent postoperative fever, or other possible manifestations such as tonsillitis, stomatitis or pneumonia. The authors did not compare the sensitivity of their method with a laboratory neutrophil count when they concluded that no cases of agranulocytosis had occurred. Agranulocytosis is typically defined as a decrease in neutrophils to less than 0.5 × 10⁹ l⁻¹. A small observational study on the diagnosis of agranulocytosis but performed on another clinical group found that of 55 cases in whom the laboratory had diagnosed agranulocytosis, 43 were detected before the occurrence of an infection, and 29 of these who were adequately treated never had any indication of infection despite the absence, or a very small number, of neutrophils.¹¹ This emphasises the weak sensitivity of clinical evaluation and underlines the importance of laboratory testing in the diagnosis of agranulocytosis.

The current fashion of ‘intellectual suggestions’ in pharmacology research threatens the very rationale of safety assessment. If a study shows only a weak health risk or indeed, none at all, despite deep flaws in its design, the conclusions are presented as sound science and objective reason, especially in medication risk assessment. This message is not new; Altman writing in the British Medical Journal commented ‘we must question whether the absence of evidence is a valid enough justification’. He also famously wrote ‘Absence of evidence is not evidence of absence’.¹²

Although metamizole as reported appears efficient and relatively well tolerated, this observational study fails to provide sufficient proof of safety to justify promoting metamizole as a first-line agent in children. And even then, the small incidence of suspected agranulocytosis cannot adequately balance the fact that it is a serious life-threatening ADR. No child need experience this unpleasant disorder when we have well known alternatives wherein the mechanism of action and the safety profile are better understood.

Acknowledgements relating to this article

Assistance with the Invited Commentary: none.

Financial support and sponsorship: none.

Conflicts of interest: none.

Comment from the Editor: this article was checked and accepted by the Editors, but was not sent for external peer-review.