The relationship between the amount of inhaled steroids delivered from pressurized metered-dose inhalers used with their recommended holding chambers and age of the patients using these devices was studied in an open randomised cross-over filter study. We recruited 1-2-month-old healthy infants (n = 21), 2-3-year-old asthmatics (n = 13), 4-6-year-old asthmatics (n = 15), and 10-15-year-old asthmatics (n = 20). Each child inhaled two puffs, administered by a single investigator, of both budesonide through Nebuchamber and fluticasone propionate through Babyhaler, on two occasions. Moreover, the 4-6-year-old group inhaled via both facemask and mouthpiece. Drug, collected on a filter interposed between holding chamber and patient, was analysed by high performance liquid chromatography. Filter dose, expressed in percent of the nominal dose, was analysed in a mixed effect linear regression model with age group, holding chamber and inhalation interface (facemask or mouthpiece) as fixed effects and subject as random effect. Filter dose from both holding chambers increased significantly with age, from 3% with Babyhaler and 7% with Nebuchamber in the youngest children, to 40-41% with both holding chambers in adolescents. Nebuchamber delivered more drug than Babyhaler (p = 0.002), but variability in drug delivery (about 11%) was similar between holding chambers. Filter dose decreased from 35% to 22% with Babyhaler, and from 42% to 27% with Nebuchamber when using a mouthpiece rather than a facemask (p < 0.0001). Delivery of inhaled steroids used with their recommended holding chambers depends from age and holding chamber, but also from the inhalation interface. Lung deposition and clinical studies comparing inhalation from holding chambers with mouthpiece and facemask are urgently required.