Sniff nasal inspiratory pressure is proposed as a noninvasive test of inspiratory muscle strength. During this manoeuvre, the nasal pressure is supposed to reflect oesophageal pressure. The aim of the present study was to compare the nasal pressure with the oesophageal pressure during a maximal sniff in children with neuromuscular disease (NM, n = 78), thoracic scoliosis (n = 12) and cystic fibrosis (CF, n = 23). A significant correlation was observed between the sniff nasal and oesophageal pressure. The ratio of the sniff nasal/oesophageal pressure was lower in patients with CF (0.72+/-0.13) than in NM patients (0.83+/-0.17) or patients with thoracic scoliosis (0.86+/-0.10). In patients with CF and NM disease, this ratio was not correlated to age or spirometric data. The difference between the sniff oesophageal and nasal pressure exceeded 15 cm H(2)O in 17, 33 and 87% of the NM, thoracic scoliosis and CF patients, respectively. Sniff nasal pressure often underestimates the strength of inspiratory muscles in cystic fibrosis. Such an underestimation occurs more rarely in neuromuscular disease disorders and thoracic scoliosis. A normal value excludes inspiratory muscle weakness but a low value requires the measurement of the oesophageal pressure.