Underestimation of airway luminal eosinophilia by quantitative sputum cytometry

On Wright-stained sputum cytospins, eosinophil differential of ≥ 1.2% is considered abnormal, and ≥ 2.3% identifies an eosinophilic endotype. We hypothesized that failure to consider free eosinophil granules (FEG), and the re-emergence (unmasking) of eosinophilia due to various reasons underestimate the prevalence of the eosinophilic endotype.

This is a retrospective analysis of our Institutional Review Board-approved clinical sputum database. Of the 24,176 examinations of sputa from patients with various airway diseases, 17,693 were viable cell counts from 9570 patients (6604 on a single occasion, 2967 from multiple occasions). The prevalence of intact eosinophil % at 1.2 and 2.3% thresholds was first examined. Then, additional evidence of eosinophilia was assessed by semi-quantitative enumeration of FEGs. In those patients whose sputa were examined on multiple occasions (at the time of an exacerbation or after corticosteroid dose was reduced), re-emergence (unmasking) of eosinophilia was assessed .

Using the threshold of eosinophilia ≥ 1.2%, 6289/17693 (35.6%) of sputa were classified as eosinophilic. This increased to 7850/17693 (44.4%) when the presence of FEGs was considered. Using the threshold of eosinophilia ≥ 2.3%, 4647/17693 (26.3%) of sputa were classified as eosinophilic. This increased to 5435/17693 (30.7%) when the presence of FEG were considered. Extrapolating from the prevalence of re-emergence observed in the 2967 patients who had sputa examined on multiple occasions to the whole sample, we estimated that eosinophilia at 1.2% threshold would be observed in at least 60% of the samples, and a clinically relevant eosinophilia at 2.3% threshold would be observed in at least 48.5% of the samples.

Using a large sputum cytometry clinical database (17,693 viable cell counts), we demonstrate that a single time point intact cell count underestimates the prevalence of eosinophilia in a variety of airway diseases. The prevalence of eosinophilia increases from 35.6 to 60% (40% underestimation) at the 1.2% threshold, and from 26.3 to 48.5% (45% underestimation) at the 2.3% clinically relevant threshold, when free granules and a second examination are considered. This has important implications to identify the eosinophilic and Th2 high endotype both for clinical trials of anti-eosinophil therapies, and to select patients who may respond well to glucocorticosteroids and anti-IL5 therapies.