We compared the effects of the most commonly used respiratory muscle (RM) training regimes: RM endurance training (RMET; normocapnic hyperpnoea) and inspiratory resistive training (IMT), on RM performance. Twenty-six healthy men were randomized into 3 groups performing 4 weeks of RMET, IMT or sham-training. Lung function, RM strength and endurance were tested before and after training. RM fatigue during intermittent hyperpnoea was assessed by twitch oesophageal (P(oes,tw)) and gastric pressures with cervical and thoracic magnetic stimulation. Respiratory sensations (visual analogue scale, 0-10) and blood lactate concentrations ([La]) were assessed during hyperpnoea. RMET increased maximal voluntary ventilation while IMT increased maximal inspiratory pressure. Both RMET and IMT increased vital capacity and RM endurance, but only RMET improved the development of inspiratory muscle fatigue (from -31% to -21% P(oes,tw)), perception of respiratory exertion (4.2+/-0.1 to 2.3+/-2.3 points) and [La] (1.8+/-0.4 to 1.3+/-0.3 mmol l(-1)) during hyperpnoea. Whether these specific RMET-induced adaptations observed during hyperpnoea would translate into greater improvements in exercise performance compared to IMT remains to be investigated.