The treatment of non-conductive olfactory disorders is to a large extent an unsolved problem. This proof-of-concept study focused on possible effects of the N-methyl-D-aspartate (NMDA) antagonist caroverine. Potential mechanisms for the hypothesized effect included reduced feedback inhibition in the olfactory bulb as a consequence of NMDA antagonistic actions and antagonism of an excitotoxic action of glutamate. A total of 77 consecutive patients with non-conductive olfactory disorders were included in the study. Fifty-one patients received caroverine for 4 weeks (120 mg/day); 26 controls matched for age, gender and duration of olfactory loss were treated with zinc sulfate for the same length of time (400 mg/day). Olfactory sensitivity was evaluated before and after treatment. Testing included assessment of n-butanol odor threshold and odor identification. When compared to baseline, treatment with caroverine improved both odor thresholds (p = 0.005) and odor identification (p = 0.042) in anosmic patients. In hyposmic patients it significantly improved odor identification ability (p = 0.041). In contrast, zinc sulfate had no significant effect on olfactory function. These results indicate that caroverine appears to be effective for the treatment of non-conductive smell disorders.