Get access

Can we predict sputum eosinophilia from clinical assessment in patients referred to an adult asthma clinic?


  • Funding: None.
  • Conflict of interest: None.



There is overwhelming evidence that asthma guidelines aimed at reducing airway inflammation are superior to those based on clinical symptoms alone. This involves targeting eosinophilic inflammation with inhaled corticosteroids.


Because induced sputum is not readily available, our study set out to investigate whether the collective or singular use of routine asthma investigations can predict sputum eosinophilia.


Eighty patients underwent skin prick testing, blood tests (IgE, full blood count), spirometry, exhaled fraction nitric oxide (FeNO), PD15 to hypertonic saline, and induced sputum testing at first assessment. A predictive model for sputum eosinophilia (defined as ≥3% eosinophils) was sought using routinely available tests.


Fifty-four subjects underwent both induced sputum and FeNO testing. Seventeen (30%) revealed eosinophilic inflammation, nine (16%) neutrophilic, four (7%) mixed granulocytic and 26 (46%) paucigranulocytic. Positive predictors for sputum eosinophilia included low forced expiratory volume in 1 s (FEV1)% predicted, raised serum eosinophil, positive smoking history, Polynesian ethnicity and negative asthma family history. There was a non-statistically significant trend for FeNO predicting sputum eosinophilia. The best combination of predictors was low FEV1% predicted, raised serum eosinophil, positive smoking history and negative family history of asthma.


This study demonstrates that the serum eosinophil count and FEV1 combined with aspects of a clinical history may provide a simple and practical alternative to assessment of airway (sputum) eosinophilia in the clinical setting. A full blood count can be performed at a substantially lesser cost and with greater accessibility than induced sputum. We feel the time has come for the clinical utility of the serum eosinophil count to be revisited.