Reliability of a rapid hematology stain for sputum cytology*

OBJECTIVE To determine the reliability of a rapid hematology stain for the cytological analysis of induced sputum samples. METHODS This was a cross-sectional study comparing the standard technique (May-Grünwald-Giemsa stain) with a rapid hematology stain (Diff-Quik). Of the 50 subjects included in the study, 21 had asthma, 19 had COPD, and 10 were healthy (controls). From the induced sputum samples collected, we prepared four slides: two were stained with May-Grünwald-Giemsa, and two were stained with Diff-Quik. The slides were read independently by two trained researchers blinded to the identification of the slides. The reliability for cell counting using the two techniques was evaluated by determining the intraclass correlation coefficients (ICCs) for intraobserver and interobserver agreement. Agreement in the identification of neutrophilic and eosinophilic sputum between the observers and between the stains was evaluated with kappa statistics. RESULTS In our comparison of the two staining techniques, the ICCs indicated almost perfect interobserver agreement for neutrophil, eosinophil, and macrophage counts (ICC: 0.98-1.00), as well as substantial agreement for lymphocyte counts (ICC: 0.76-0.83). Intraobserver agreement was almost perfect for neutrophil, eosinophil, and macrophage counts (ICC: 0.96-0.99), whereas it was moderate to substantial for lymphocyte counts (ICC = 0.65 and 0.75 for the two observers, respectively). Interobserver agreement for the identification of eosinophilic and neutrophilic sputum using the two techniques ranged from substantial to almost perfect (kappa range: 0.91-1.00). CONCLUSIONS The use of Diff-Quik can be considered a reliable alternative for the processing of sputum samples.