February 2026

In both patient and mice liver biopsies, APAP-exposed livers showed significantly more eosinophils per high-powered field (eos/hpf) than unexposed, healthy controls. Interestingly, the authors showed that liver injury, measured by ALT and percentage of necrosis, was more severe in ΔdblGata1 eosinophil-knockout mice compared with wild-type mice—a pattern that seemed to be rescued by acute transfer of eosinophils. The investigators then pursued a series of experiments to isolate the pathway responsible for eosinophil recruitment. Using flow cytometry in both Il-33-/- and macrophage-deficient mice, they showed that hepatic eosinophil recruitment was dependent on Il-33 in conjunction with macrophages. Experiments with an Il-33-GFP reporter mice revealed that liver sinusoidal endothelial cells (LSECs) were critical in secreting Il-33, while depleting macrophages led to a steep drop-off in serum levels of CCL24, a known eosinophil chemoattractant, instead. Treating macrophages in vitro with Il-33 did not lead to CCL24 release, indicating that another cell type may be involved. When cultured together with eosinophils derived from APAP-treated mice, the authors found that these eosinophils, which expressed the receptor for Il-33, ST2, were required for Il-33-dependent CCL24 secretion from macrophages.

This study helped significantly advance our understanding of the normal homeostatic role of eosinophils. Xu et al. brilliantly demonstrated that eosinophils can play a protective role and that recruitment in APAP-induced liver injury is dependent on Il-33, released by LSEC, and cross-talk between ST2+ eosinophils and CCL24-secreting macrophages. Strengths of the study included its comprehensiveness, utilization of multiple models of lung injury, and use of both in vivo and in vitro models. However, the authors have yet to demonstrate a therapeutic effect of eosinophil recruitment. Future studies should seek to explore potential downstream mechanisms of how eosinophils help to prevent liver injury.

In this study, investigators developed a machine learning–based prediction model using an open claims database including patients (with/without HES ICD-10 code) who had ≥2 blood eosinophil counts >1,000 cells/μL. This lower threshold was chosen to capture a broader population, including patients with treatment-suppressed counts. Selected predictors included demographics, disease manifestations, comorbidities, treatments, and healthcare utilization.

Selected variables were incorporated into a statistical model to estimate  predicted probabilities of HES risk across thresholds from 0.1–0.9. Key predictors included having a bone marrow biopsy, blood eosinophil count >3,000/µL, and history of asthma. Model performance was evaluated using Receiver Operating Characteristics  Area under the curve (ROC AUC) and Precision-Recall AUC (PR AUC) to assess overall discrimination and precision, both demonstrated strong results (ROC AUC = 0.82; PR AUC = 0.83). At a 0.7 probability threshold, 6,233 additional patients with predicted HES were identified and exhibited characteristics similar to diagnosed cases. The estimated model-derived prevalence of total HES (diagnosed plus predicted) was 5.65% among individuals with elevated eosinophil counts, suggesting underdiagnosis.

Strengths of this study include a large real-world dataset and a robust predictive approach. Limitations include potential misclassification of HES in secondary datasets and limited generalizability to patients with lower blood eosinophil counts. Given that HES is often diagnosed late, this model may enable earlier identification of HES patients, prompting timely referral & intervention, although it is not a stand-alone diagnostic tool and requires clinical interpretation and further validation. Rather than distinguishing primary (clonal/idiopathic) HES from secondary (reactive) eosinophilia, the model is designed to flag patients whose overall patter of findings resemble those with a recorded HES diagnosis. In addition to external validation of model and extension to identification of tissue-based HES, future studies could be focused on developing machine-learning based tools to facilitate earlier diagnosis of HES in the population. Overall, these findings suggest that machine learning-based prediction may improve HES detection and reveal an underrecognized disease burden.