Sputum analysis is a useful technique for the study of airway inflammation. In asthma, dithiothreitol (DTT) is used to disperse cells from surrounding mucus; however, the applicability of these processing methods to cystic fibrosis (CF) sputum is unknown. In order to compare two methods for processing sputum of patients with CF, sputum was obtained from 11 subjects with CF (8 female, aged 9-21 years). The sample was split into 2 portions and sputum dispersal using DTT was compared with an enzyme mixture (E) of deoxyribonuclease, hyaluronidase, and galactosidase. Outcomes assessed were sample quality, cell viability (percent cells excluding trypan blue), total cell count (TCC), neutrophil count, and elastase immunoreactivity (percent cells positive). Sample quality (enzymes vs. DTT, 8.3 +/- 0.3 vs 7.6 +/- 0.4, mean +/- SEM) and cell viability (enzymes vs. DTT, 75.0% vs. 68.0%, median) were similar for both methods. Sputum total cell count (20.5 x 10(6)/ml vs. 12.0 x 10(6)/ml, median; P = 0.01) and neutrophil count (13.4 x 10(6)/ml vs. 5.5 > 10(6)/ml, median; P = 0.02) were significantly higher with E. Elastase immunoreactivity was lost after processing with E (19.0% vs. 39.5%, median; P = 0.04). When purified peripheral blood neutrophils were incubated with DTT and E, there was no reduction in neutrophil viability, suggesting that the reduced neutrophil number in CF sputum was not due to a toxic effect of DTT but rather incomplete dispersal. We conclude that published sputum processing methods for asthma using DTT give false results when applied to CF sputum, which should be processed using an enzyme mixture.