An understanding of mechanisms, potential benefits and risks of antidotes is essential for clinicians who manage poisoned patients. Of the dozens of antidotes currently available, only a few are regularly used. These include activated charcoal, acetylcysteine, naloxone, sodium bicarbonate, atropine, flumazenil, therapeutic antibodies and various vitamins. Even then, most are used in a minority of poisonings. There is little randomized trial evidence to support the use of most antidotes. Consequently, decisions about when to use them are often based on a mechanistic understanding of the poisoning and the expected influence of the antidote on the patient's clinical course. For some antidotes, such as atropine and insulin, the doses employed can be orders of magnitude higher than standard dosing. Importantly, most poisoned patients who reach hospital can recover with supportive care alone. In low risk patients, the routine use of even low risk antidotes such as activated charcoal is unwarranted. In more serious poisonings, decisions regarding antidote use are generally guided by a risk/benefit assessment based on low quality evidence.