The objective of this investigation was to study the relation between size and position of a mask leak on spacer output and lung dose. An upper-airway model (SAINT model, Erasmus MC) was connected to a breathing simulator. Facemasks with leaks ranging between 0 and 1.5 cm(2) were examined. Leaks were located close to the nose or close to the chin. During simulated breathing, 200 microg budesonide (Pulmicort, AstraZeneca) was delivered to the model via NebuChamber (AstraZeneca) with facemask. Spacer output and lung dose were measured by placing a filter between spacer and facemask or between model and breathing simulator, respectively. Budesonide trapped on the filter was quantified by means of HPLC, and expressed as percentage of the nominal dose. Mean spacer output doses for the nose position were 50, 38, 28, 12, 10, 6, and 0%, and for the chin position were 50, 40, 31, 11, 9, 4, and 0% for leaks of 0, 0.05, 0.1, 0.16, 0.2, 0.3, and larger than 0.4 cm(2), respectively. Mean lung doses for the nose position were 10, 8, 6, 3, 3, 1, 0, 0, 0, and 0%, and for the chin position were 10, 9, 8, 6, 6, 5, 1, 1, 0, and 0% for leaks of 0, 0.05, 0.1, 0.16, 0.2, 0.3, 0.4, 0.5, 1, and 1.5 cm(2). Efficiency of a pMDI-spacer facemask strongly depends on the size of a facemask leak. Spacer output did not depend on the position of the leak. Lung dose was higher for leaks near the chin than for leaks near the nose.