Skip to main content
main-content

15-05-2017 | Asthma | Highlight | Article

Editor's pick

Multivariate model predicts eosinophilic asthma without sputum induction

medwireNews: Researchers have designed a multivariate prediction model for identifying eosinophilic asthma that uses the activation state of peripheral blood eosinophils rather than sputum analysis.

The model includes the established asthma measurements of blood eosinophil count, fractional exhaled nitric oxide (FeNO), Asthma Control Questionnaire (ACQ) score, medication use, nasal polyposis, and aspirin sensitivity combined with six measures of peripheral blood granulocyte activation status.

The researchers stress in Allergythat the responsiveness of neutrophils and eosinophils to stimulation with formyl-methionyl-leucyl phenyalanine was “essential to come to a sensitive diagnostic test and adds to the ongoing scientific debate about the biological relevance of granulocyte responsiveness in asthma.”

They report that the sensitivity of the prediction model increased from 47.6% to 90.5% with the addition of granulocyte responsiveness.

The 12 parameters included in the model were identified from a total of 26 using nonlinear principal component analysis. These parameters explained the greatest variance among 115 patients with asthma recruited from the University Medical Centre in Utrecht, the Netherlands.

The combined scoring and loading plot for the patients showed a distribution that was largely determined by markers of eosinophilic inflammation.

Generally, the eosinophilic asthma phenotype was characterized by increased FeNO levels and ACQ scores, nasal polyposis, aspirin sensitivity, and high treatment intensity, with a high blood eosinophil count and cells refractory to stimulation.

By contrast, patients with a neutrophilic or paucigranulocytic phenotype had lower FeNO levels and ACQ scores, low treatment intensity, and did not have nasal polyps or aspirin sensitivity. Their eosinophil blood counts were low and responsive to active stimulation.

This predictive model correctly identified sputum eosinophilia in the original 115 patients with a sensitivity of 90.5% and a specificity of 91.5%. When a validation cohort of 34 patients with asthma from the UK was tested sputum eosinophilia was correctly identified with 76.9% sensitivity and 71.4% specificity.

Bart Hilvering (University Medical Centre, Utrecht, the Netherlands) and colleagues believe that the relatively low sensitivity and specificity in the UK cohort is likely explained by the high percentage of patients in this group taking oral corticosteroids, at threefold that of the Utrecht cohort.

“Oral corticosteroids are known to induce apopotosis in eosinophils,” they explain. “Therefore, these patients are particularly less likely to have sputum eosinophilia, leading to the ‘false’ conclusion that they do not suffer from eosinophilic asthma,” they add.

They were able to demonstrate this by excluding the UK cohort patients taking oral corticosteroids, which improved the accuracy of the model from 73.5% to 79.2%. The sensitivity fell to 72.7% and the specificity increased to 84.6%.

Therefore, the model “identified an important group of patients with potentially eosinophilic inflammation that rendered noneosinophilic in sputum most likely due to [oral corticosteroid] use,” the researchers comment.

They therefore suggest that “the prediction model […] is more suitable for asthma classification in patients not on [oral corticosteroids].”

By Lucy Piper

medwireNews is an independent medical news service provided by Springer Healthcare. © 2017 Springer Healthcare part of the Springer Nature group

Related topics