TOPLINE

Children with asthma had higher serum levels of eosinophil-derived neurotoxin, a protein released by activated eosinophils, than children with other allergic conditions or healthy children. Blood eosinophil counts were also higher in children with asthma than in healthy children, but the difference between the asthma and allergic groups was not statistically significant.

METHODOLOGY

• Researchers conducted a prospective study using real-world data to determine whether serum eosinophil-derived neurotoxin and blood eosinophil count could help distinguish children with asthma from those who had other allergic diseases or healthy children.
• They included 74 children with asthma (an active diagnosis for at least 1 year), 35 with other allergic conditions but no asthma or recurrent wheezing, and 32 healthy children. Median ages ranged from 8 to 11.5 years across groups, and 53%-63% of children were girls.
• Blood samples were analyzed for eosinophil-derived neurotoxin, blood eosinophil count, and neutrophil count. Children with asthma also completed an asthma control test questionnaire and, if older than 6 years, an optional fractional exhaled nitric oxide test.
• Asthma was classified as type 2 (T2)-high endotype if blood eosinophil count exceeded 0.3×10⁹ cells/L or fractional exhaled nitric oxide exceeded 25 parts per billion, and T2-low otherwise.
• Researchers developed classifier models using machine learning (the XGBoost algorithm) to evaluate the diagnostic contributions of eosinophil-derived neurotoxin and blood eosinophil count, incorporating demographic factors, diagnosis codes for allergic conditions other than asthma, and blood biomarkers.

TAKEAWAY

• The median serum levels of eosinophil-derived neurotoxin were 55.5 µg/L in children with asthma, 40.7 µg/L in those with other allergic conditions, and 30.1 µg/L in healthy children. Concentrations were significantly higher in the asthma group than in both the healthy (P < .001) and allergic groups (P = .0238).
• The median blood eosinophil count was significantly higher in children with asthma than in healthy children (0.29 × 10⁹ cells/L vs 0.16 × 10⁹ cells/L; P = .002), but the difference between the asthma and allergic groups was not statistically significant. Neutrophil counts did not differ significantly across the three groups.
• Among children with asthma, those with T2-high endotype had significantly higher levels of eosinophil-derived neurotoxin than those with T2-low endotype (P < .001).
• In classifier models based on machine learning, adding either eosinophil-derived neurotoxin or blood eosinophil count as a feature raised overall accuracy from 0.522 to 0.565 for classifying children into asthma, allergic, or healthy groups; both markers ranked among the important features.

IN PRACTICE

"In this study, we have shown that serum EDN [eosinophil-derived neurotoxin] levels are highest in children with asthma, intermediate in children with other allergies, and lowest in healthy children. Serum EDN may increase the performance of machine learning-based diagnostic classifiers for asthma," the authors wrote.

SOURCE

Yifei Yang, PhD, with Mayo Clinic, Rochester, Minnesota, was the corresponding author of the study, which was published online on October 5 in The Journal of Allergy and Clinical Immunology: Global.

LIMITATIONS

The researchers lacked data on children's patterns of allergen sensitization. The test set was small, with only three healthy children, limiting comparisons of accuracy between machine learning models. The study took place at a single center, and most children were non-Hispanic White, limiting the generalizability of the findings to other populations.

DISCLOSURES

The study was supported by Mayo Clinic and Thermo Fisher Scientific. Two authors reported being employees of Thermo Fisher Scientific.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.