Stopping trials early may result in insufficient evidence being accrued

EDITOR—Fan et al's paper is the second in the BMJ recently to investigate prophylaxis against adverse effects of snake antivenom.1 Earlier, Premawardhena et al in Sri Lanka concluded that giving low dose subcutaneous adrenaline before antivenom could prevent acute adverse reactions.2 The Brazilian study (by Fan et al) showed no benefit in using promethazine as pretreatment.1 Perhaps more interesting than the conclusions, however, is that both these clinical trials were stopped early because of interim analyses; they provide examples of the use and pitfalls of such analyses.

The Sri Lankan group stopped recruitment halfway through its …