Background: Induced sputum differential cell counts have been advocated as a method of noninvasively assessing airway inflammation in asthma and other airway diseases. Relatively little is known about the between-observer repeatability of sputum differential cell counts and the factors that influence it.
Objective: To assess the between-observer variability of induced sputum cell counts.
Methods: Sputum was induced and processed using standard techniques. Forty-two slides from 38 patients (31 with asthma, seven normal subjects) were randomly selected. Slides were classified as good (<20% squamous cells and >50% viability; n = 24); low viability (<50% viability; n = 10) and high squamous cell contamination (>20% squamous cells; n = 8). Two blinded observers counted between 200 and 400 nonsquamous cells and agreement was assessed by the intraclass correlation coefficient (ICC) and the standard deviation of between-observer differences (SD).
Results: The overall ICC were 0.9, 0.89, 0.9 for eosinophils, neutrophils and macrophages and 0.29 and 0.69 for lymphocytes and epithelial cells. Repeatability was greater in slides classified as good compared with slides with low cell viability and particularly excess squamous cell contamination.
Conclusions: We have shown that the overall between-observer repeatability of the differential eosinophil, neutrophil and macrophage cell counts is good. Low cell viability and particularly excess squamous cell contamination reduce between-observer repeatability suggesting that techniques that ensure high cell viability and reduce squamous contamination would be an advantage.