The need for careful testing of new drugs in animal models before study in humans has been recognised by physicians since the First World War. Now, first human studies on new drugs are subject to detailed government guidelines, which in the European Union are presently being reinforced through the wide-ranging Clinical Trials Directive. However, despite their long history and widespread application, these guidelines are empirical and have been formulated with a paucity of critical scientific evidence. Here, we review the principles and the available, albeit limited, evidence that support the design and conduct of preclinical studies in a way that permits effective and safe first-dose studies of potential new medicines in humans.