Methods: We randomly allocated 158 mothers of food-allergic children prescribed an EAI for the first time, to Anapen or Epipen with appropriate training. Treatment was allocated by a third party using a computer-generated randomisation sequence. The primary outcome, ability to effectively administer epinephrine in a recorded simulated anaphylaxis scenario, was assessed by a paediatric allergist independent of the trial and host institution. ISRCTN12504076
Results: We evaluated ability to administer epinephrine at 1 year in 110/158 (70%) randomised participants. Overall success rates were low, and similar between the 2 groups - 28/51 (55%) using Anapen, and 35/59 (59%) using Epipen (P=0.64). The primary reason for failure differed between groups (P<0.001). Failure to remove all safety caps was most common for Anapen (33%); using the wrong end of the device for Epipen (17%). When actual device-specific epinephrine delivery times were taken into account, success rates were 30/51 (59%) for Anapen and 42/59 (71%) for Epipen (P=0.17). Digital injection - the only adverse event recorded – occurred in 8/59 (14%) in the Epipen group, and 0/51 (0%) in the Anapen group (P=0.007).
Conclusions: EAI success rates were similar for Anapen and Epipen 1 year after training, but digital injection was more common with Epipen. Optimal EAI design should include a single safety cap, and easy identification of the needle end.