Background Modern gas analysis techniques permit real time and on-line quantification of multiple volatile trace gases within a single exhalation. However, the influence of various respiratory manoeuvres affecting exhalation flow and the kinetics of metabolite release to the gas-phase remain largely unknown.

Methods We examined variation in the concentrations of selected trace gases over a range of expiratory flows (50; 100; 250 ml/s) and after 30 second periods of breathold and paced hyperventilation. On-line measurement of breath samples from healthy volunteers (n=10) was performed by proton transfer mass spectrometry.

Results Exhaled acetone increased with higher expiratory flow rate (805, 838, 898 ppb, p=0.02). Levels of methanol (206 vs 179 ppb, p<0.01), acetaldehyde (26 vs 22 ppb, p<0.01), ethanol (410 vs 208 ppb, p=0.01) and dimethyl sulphide (113 vs 103 ncps, p<0.01) fell significantly following 30s hyperventilation. After 30 second breathold levels of methanol (206 vs 217 ppb, p=0.02), acetone (805 vs 869 ppb, p<0.01), isoprene (348 vs 390 ppb, p=0.02) and dimethyl sulphide (113 vs 136 ncps, p=0.02) increased significantly. Variation in respiratory parameters did not significantly alters the level of acetonitrile, propanol and butyric acid within the breath of healthy subjects.

Conclusions These findings demonstrate that respiratory manoeuvres significantly influence the measured concentration of a number of exhaled VOCs that are of potential importance within the clinical setting. Our results support the adoption of standardised practices for breath gas analysis by on-line and real time mass spectrometry methods.