Abstract

Spirometric variables were obtained from nine symptomless subjects, who performed forced expiratory manoeuvres with three Vitalograph spirometers at three different ambient temperatures (36.5 degrees C, 24.1 degrees C, and - 7.3 degrees C) over three days according to a Latin square design. Analysis of variance showed no significant difference between values at different ambient temperatures when measured at ATPS. Correction of the results to BTPS resulted in a significant difference in values at the three ambient temperatures for all measurements other than the Vmax25. The mean error introduced by conversion to BTPS varied from 7% (SD 5%) for the FVC at 24.1 degrees C to 30% (21%) for Vmax50 at - 7.3 degrees C. Possible explanations for these observations include the increasing compliance of the spirometer bellows with increasing temperature, relatively slow cooling of gases within the spirometer, and a combination of these effects. We conclude that there is no evidence that conversion to BTPS improves the accuracy of measurements made on a Vitalograph spirometer. Further studies on other spirometers are needed to see whether conversion to BTPS is always appropriate.