The relationship between exposure to moclobemide and the occurrence of adverse events (AEs) was quantified by applying a population approach to data collected during phase III/IV studies. A total of 965 measurements of moclobemide concentration in plasma were available from 357 depressed patients included in a series of clinical studies; 539 AEs were reported by the 236 patients with at least one concentration measurement > or = 400 micrograms/liter. Using relatively few blood samples per patient (mean, 2.7), the average concentration in plasma on the day an AE was observed was calculated (CAE,pred) from population kinetic parameter estimates. The frequency of AEs, normalized by exposure time in months (AEs/Mo), was grouped according to dose and concentration categories and examined in relation to age, gender, adverse event severity, and concomitant medication. CAE,pred was more powerful than the observed concentration or daily dose in detecting relationships between exposure to moclobemide and the occurrence of AEs. There was a relationship between CAE,pred and AEs/Mo, with 0.58 events at concentrations between 1,000 and 1,999 micrograms/liter and 2.78 events at concentrations of 4,000 micrograms/liter and above. Within each concentration category, the number of AEs decreased with increasing severity. Advancing age was not associated with an increase of AEs/Mo. AEs were more frequent in females than in males in all treatment intensity categories (ratio, 1.4). There was no indication of supra-additive effects by concomitant medication, including all drugs used in the study population. This investigation demonstrates the successful use of observational data from clinical studies in the assessment of quantitative relationships between drug exposure and the occurrence of AEs.