BACKGROUND Diaphragm strength can be assessed by the measurement of gastric (TW PGA), oesophageal (TW POES), and transdiaphragmatic (TW PDI) pressure in response to phrenic nerve stimulation. However, this requires the passage of two balloon catheters. A less invasive method of assessing diaphragm contractility during stimulation of the phrenic nerves would be of clinical value. A study was undertaken to determine whether pressure measured at the mouth (TW PM) during magnetic stimulation of the phrenic nerves accurately reflects TW POES, and to investigate the relations between TW PM and TW PDI; and also to see whether glottic closure and twitch potentiation can be avoided during these measurements. METHODS Eight normal subjects and eight patients with suspected respiratory muscle weakness without lung disease were studied. To prevent glottic closure magnetic stimulation of the phrenic nerves was performed at functional residual capacity during a gentle expiratory effort against an occluded airway incorporating a small leak. TW PDI, TW POES, and TW PM were recorded. Care was taken to avoid potentiation of the diaphragm. RESULTS In normal subjects mean TW PM was 13.7 cm H2O (range 11.3-16.1) and TW POES was 13.3 cm H2O (range 10.4-15.9) with a mean (SD) difference of 0.4 (0.81) cm H2O. In patients mean TW PM was 9.1 cm H2O (range 0.5-18.2) and TW POES was 9.3 (range 0.7-18.7) with a mean (SD) difference of -0.2 (0.84) cm H2O. The relation between TW PM and TW PDI was less close but was well described by a linear function. In patients with diaphragm weakness (low sniff PDI) TW PM was < 10 cm H2O. CONCLUSIONS TW PM reliably reflects TW POES and can be used to predict TW PDI in normal subjects and patients without lung disease. TW PM may therefore be a promising non-invasive, non-volitional technique for the assessment of diaphragm strength.