Retention of mouth-to-mouth, mouth-to-mask and mouth-to-face shield ventilation techniques is poorly understood.A prospective randomised clinical trial was undertaken in January 2004 in 70 candidates randomly assigned to training in mouth-to-mouth, mouth-to-mask or mouth-to-face shield ventilation. Each candidate was trained for 10 min, after which tidal volume, respiratory rate, minute volume, peak airway pressure and the presence or absence of stomach inflation were measured. 58 subjects were reassessed 1 year later and study parameters were recorded again. Data were analysed with ANOVA, chi(2) and McNemar tests.Tidal volume, minute volume, peak airway pressure, ventilation rate and stomach inflation rate increased significantly at reassessment with all ventilation techniques compared with the initial assessment. However, at reassessment, mean (SD) tidal volume (960 (446) vs 1008 (366) vs 1402 (302) ml; p<0.05), minute volume (12 (5) vs 13 (7) vs 18 (3) l/min; p<0.05), peak airway pressure (14 (8) vs 17 (13) vs 25 (8) cm H(2)O; p<0.05) and stomach inflation rate (63% vs 58% vs 100%; p<0.05) were significantly lower with mouth-to-mask and mouth-to-face shield ventilation than with mouth-to-mouth ventilation. The ventilation rate at reassessment did not differ significantly between the ventilation techniques.One year after a single episode of ventilation training, lay persons tended to hyperventilate; however, the degree of hyperventilation and resulting stomach inflation were lower when a mouth-to-mask or a face shield device was employed. Regular training is therefore required to retain ventilation skills; retention of skills may be better with ventilation devices.