Introduction
Endotyping is important for risk assessment of asthma, yet studies have primarily used adult populations and/or protocolled assessments that may be unavailable in daily clinical care. The value of routine data collection for childhood asthma endotyping remains largely unexplored.

Methods
Using Danish nationwide databases, 29,851 children aged 2-17 years with ICS-treated asthma in 2015 were identified. Impact of asthma outcomes was investigated based on traits available in daily clinical care: blood eosinophils (EOS) >300 cells/µL, atopy (positive RAST or elevated total IgE levels), and non-atopic traits identified by maternal smoking and viral infections.

Results
Most children (61.8%) were atopic or had elevated EOS, whereas 32.2% had a mixed endotype with both non-atopic traits and elevated EOS or atopy, with the latter being more likely to have severe asthma, and experience poor control and exacerbations.

Non-atopic traits were associated with increased odds of poor asthma control (maternal smoking OR 1.42 (1.30-1.58), viral infections 1.59 (1.48-1.71). The impact of maternal smoking on poor control diminished with increasing child age.

Atopic traits and EOS were associated with both asthma exacerbations and asthma severity (ORs 1.55 (1.38-1.73) and 1.42 (1.24-1.63) for eosinophils, 1.41 (1.20-1.66) and 1.31 (1.08-1.60) for atopy). For eosinophils, a dose-dependent response was seen in terms of risk of poor asthma outcomes.

Conclusion
Endotype clusters can be reliably replicated in real-world datasets utilizing routinely collected data from daily clinical care and are predictive of disease burden. As such, already available data can help identify children at risk for adverse asthma outcomes.