Culprit allergen prevalence in polyreactive individuals reflects prescription trends: A tool for estimating comparative anaphylaxis risk using the example of neuromuscular blocking agents.

Prescription-event monitoring (PEM) is the current gold standard for determining the risk of rare drug side-effects and comparing the risk between agents; however, spontaneous or prompted reporting schemes have low case-detection rates and exposure may be difficult to estimate. A novel method is described that allows a comparative adverse event rate between two drugs to be estimated-based on patterns of cross-reactivity-requiring only a sample of cases and no direct knowledge of drug exposure rates. Agreement was compared between the novel method and historical estimates of risk using PEM for comparative risk of rocuronium versus vecuronium anaphylaxis. The novel method was applied to a sample of patients investigated by the Western Australian Anaesthetic Drug Reaction Clinic over a 21-year period. Relative population exposure was estimated from the number of patients with either rocuronium or vecuronium anaphylaxis subsequently shown to be reactive on skin testing to both agents. This was used to correct the total number of cases of hypersensitivity triggered by each agent. Measures of spread were by bootstrap sampling. Historical estimates were gathered by literature review. Additional comparisons of agreement between estimates made by the novel method and PEM were made using cross-reactivity data and PEM rates reported in the literature. There was agreement between estimates of comparative anaphylaxis risk between the novel method and PEM. Two-hundred and twenty-eight cases of anaphylaxis were observed, 89% caused by rocuronium. Patients reactive to both agents were more likely to be female, and had a higher acute mast cell tryptase level. Patients with a history of rocuronium anaphylaxis were more likely to be reactive to one agent only (69% . 33%,  < 0.01). It was estimated that rocuronium was prescribed 3.9 times more frequently than vecuronium. When the observed proportion of cases was corrected for exposure rate, the risk of rocuronium anaphylaxis was 2.2 times that of vecuronium (95% confidence interval 1.7 to 2.8). The median risk from historical estimates was 4.7 times, while the previous PEM estimate in Western Australian was 3.0 times. Using a subgroup of patients susceptible to the same side-effect of two drugs, the relative exposure rate and corrected comparative risk of an adverse effect can be estimated for a population. Using this technique, which requires assessment only of cases to estimate relative exposure rates, we have estimated that the risk of anaphylaxis from rocuronium to be 2.2 times that of vecuronium in Western Australia.