(338) Risk of agranulocytosis with metamizole in comparison with alternative medications based on health records in Spain

Background: Agranulocytosis is a rare and severe event associated with metamizole.

Objectives: We aimed to estimate the risk of agranulocytosis with the use of metamizole compared to non-steroidal anti-inflammatory drugs (NSAID).

Methods: A cohort study of new users (1st dispensation after ≥6 months without prescription) of metamizole versus NSAID was performed with BIFAP (Pharmacoepidemiologic Research Database in Public Health Systems; Spain). Patients were followed
from the day after the 1st dispensation of metamizole or NSAID up to a recorded hospitalized idiosyncratic agranulocytosis or neutropenia or end of their first continuous treatment period (>30 days without supply). Manual review of clinical records around the hospital date was performed. Incidence rate (IR) and adjusted hazard ratios (HR) of agranulocytosis with metamizole versus NSAID were estimated using a lag-time of 7 days after 1st dispensation to avoid treatment probably indicated for agranulocytosis symptomatology (reverse causation). In secondary analyses we used 1) opioids-paracetamol as negative control, 2) any hospitalized neutropenia as outcome or 3) all agranulocytosis 1 day after 1st dispensation (i.e. without lag-time).

Results: The cohorts included 444,972 new users of metamizole, 3,814,367 NSAID and 3,129,221 opioids-paracetamol. IR of agranulocytosis was 8.52 (N=3), 1.62 (N=5) and 3.72 (N=13) /10^7 person-weeks, respectively. Two (67%), 0 and 2 (15%) of cases had neoplasia, respectively. HR of agranulocytosis 7 days after 1st dispensation with metamizole was 4.40 [95% CI: 0.90-21.57] versus NSAID and 2.45 [0.68-8.83] versus opioid-paracetamol. HR of neutropenia 7 days after 1st dispensation with metamizole was 2.98 [1.57-5.65] versus NSAID. HR of agranulocytosis 1 day after 1st dispensation resulted 7.20 [1.92-26.99] versus NSAID.

Conclusions: This study confirmed that agranulocytosis is a very rare outcome. It was more frequent with metamizole than other analgesics, although this could not be statistically established in main analysis due to insufficient sample size. The higher hazard with metamizole was only significant when including also all agranulocytosis hospitalised during the first week of treatment. That could include cases of drug-induced agranulocytosis but also cases initiated before the decision to treat (reverse causality). The precise onset of the symptoms of agranulocytosis leading to analgesic treatment is paramount to elucidate that association, as well as the clarification of metamizole-induced agranulocytosis mechanism. The link with neutropenia supported the implication of metamizole, while unmeasured triggers (ex. cytotoxic medication) may still be present.