Standardization of conditions for peak expiratory flow (PEF) monitoring is much more difficult in practice than for laboratory spirometry. Patients are usually asked to record PEF before medication. The aim of this study was to determine the effect of prior bronchodilator use on PEF outcome measures in a clinical trial. Electronic PEF records from 43 subjects with poorly controlled asthma were examined to determine the frequency with which beta2-agonist was inhaled <4 h before PEF measurement, as such PEF are potentially "postbronchodilator". The effect of inclusion of such PEF values on improvement in PEF outcome measures after 8 weeks of inhaled budesonide was calculated. Subjects were asked to record PEF before medication. During run-in, the median frequency of postbronchodilator PEF was 29%, falling to 0% after 8 weeks of budesonide. Inclusion of postbronchodilator PEF led to an overestimation of average morning, evening and daily PEF during run-in (p<0.001). Improvement in these indices with treatment was, therefore, underestimated. Minimum morning PEF expressed as per cent personal best was unaffected. Subjects may not be able to withhold beta2-agonist for 4 h before every peak flow reading. This may change as the level of asthma control changes, leading to a systematic bias in clinical trial end-points or inaccuracy in individual treatment decisions. Simple changes to peak expiratory flow instructions and analysis are proposed.