We investigated the sensitivity and reproducibility of a test procedure for measuring hydrogen peroxide (H2O2) in exhaled breath condensate and the effect of storage of the condensate on the H2O2 concentration, and compared the results to previous studies. Twenty stable COPD patients breathed into our collecting device twice for a period of 10 min. The total exhaled air volume (EAV) and condensate volume were measured both times and the H2O2 concentration of the condensate was determined fluorimetrically. The concentration was measured again after freezing the reaction product at −70°C for a period of 10, 20 and 40 days. We collected 2–5 ml condensate in 10 min. The EAV and condensate volumes were strongly correlated. There was no significant difference between the mean H2O2 concentration of the first and second test. We obtained a detection limit for the H2O2 concentration of 0·02 μ moll−1. The H2O2 concentration appeared to remain stable for a period up to 40 days of freezing. Compared to previous studies we developed a more efficient breath condensate collecting device and obtained a lower H2O2 detection limit. The measurement of exhaled H2O2 was reproducible. In addition, storage of the samples up to 40 days showed no changes in H2O2 concentration.