A problem with short recorded lengths of expiration, encountered in a large heart attack intervention trial, illustrates the importance of standardization and training in spirometric lung function testing. At baseline, half of the trial's clinical centers had mean FVC values that were between 500 and 1,100 ml below those predicted for these centers. To quantify the effects of underrecorded forced expirations on the FVC, the FEV1/FVC%, the FEF70-80%, the FEF 25-75%, and the FEF45-55%, a study of 80 complete spirograms of good quality was undertaken. The findings were that short lengths of expiration cause underrecording of the FVC, inflating all of the examined FVC-dependent spirometric indexes. The inflation was fairly uniform across all obstructions for the forced expiratory flow rates, but it increased markedly with level of obstruction for the FEV1/FVC%. Virtually no subjects exhibited low values of these spirometric indexes after 3 s of expiration, and the number of subjects with low values was still substantially underestimated after 6 s of expiration. Whereas 18 (23%) of the subjects had FEV1/FVC% greater than or equal to 80% and 21 (26%) had ratios less than or equal to 69% based on complete expiration, 60 (75%) of the subjects had ratios greater than or equal to 80% and only 2 (3%) had ratios less than or equal to 69%, after 3 s of expiration. Even after 6 s, 26 subjects (33%) had ratios greater than or equal to 80%, and only 10 (13%) had ratios less than or equal to 69%. Therefore, completely recorded expirations are essential for accurate measurement of the FVC-dependent spirometric indexes.