Savoxepine, an atypical neuroleptic compound developed in the 19805, was believed to act via selective limbic dopamine D-2-receptor blockade. The results of the presented double-blind, randomised, controlled clinical trial comparing savoxepine (n = 58) with haloperidol (n = 29) did not confirm the preclinical data suggesting that savoxepine would produce fewer extrapyramidal symptoms than the comparator. Animal data and PET-results obtained a posteriori suggested that this unfavourable outcome may have been due to the conventional, step-wise dose increase strategy used in the study leading to a high dopamine D-2-receptor occupancy in the striatum thus eliciting EPS. Using either a slower dose-titration or a high, single loading dose followed by a low maintenance dosing may have offered the possibility to obtain a good antipsychotic effect together with low incidence of EPS. In future clinical trials with new neuroleptics, the preclinical data should be carefully evaluated, and drug brain kinetic parameters taken into consideration.