BACKGROUND Objective measurement of lung function is considered essential in the management of patients with asthma and COPD. Many primary care practitioners lack the means necessary to obtain these measurements conveniently. To meet this need, electronic spirometers, offering portability, ease of operation, and timesaving readout options have been introduced. We compared the accuracy of a typical pneumotachograph-based device with a conventional volume displacement spirometer. METHODS We compared indexes of pulmonary function (FVC, FEV1, mean forced expiratory flow during the middle half of FVC, [FEF25-75%], and peak expiratory flow rate [PEFR]) measured by the handheld device with those measured by a conventional spirometer in 75 white subjects (33 men, 42 women) with a median age of 43 years (22 to 77 years) who were either healthy or were referred to the pulmonary function laboratory of a large tertiary care teaching hospital. The order of the instrument tested first was randomized and the patients were blinded to which instrument was being studied. RESULTS There was a linear relationship between instruments for all indexes measured (r = 0.97, 0.98, 0.94, 0.94 for FVC, FEV1, FEF25-75%, and PEFR, respectively, for all p < 0.001). The random error (precision) was within 5% only for FEV1. The mean of the differences between the values obtained using both instruments (the bias) +/- limits of agreement (+/- 2 SD) were 0.06 +/- 0.56 L for FVC (p = NS), 0.2 +/- 0.44 L for FEV1 (p < 0.05), 0.61 +/- 1.26 L/s for FEF25-75% (p < 0.05), and 0.44 +/- 1.9 L/s for PEFR (p < 0.05). CONCLUSION Our data suggest that measurements obtained using the pneumotachograph device are closely related to those obtained by volume displacement spirometry and that the handheld device may be useful in clinical practice. However, because the limits of agreement are wide and the difference between the two instrument measurements are significant for FEV1, FEF25-75%, and PEFR, the bias between them is not consistent nor is it insignificant. Therefore, the measurements made with the two types of machine cannot be used interchangeably.