BACKGROUND Chronic graft-versus-host disease (cGVHD) is the most frequent late complication after allogeneic hematopoietic stem cell transplantation (allo-HSCT) and is associated with non-relapse mortality. If recipients with cGVHD are not properly treated at the appropriate time, cGVHD can lead to severe complications. Therefore, early prediction of cGVHD is required to improve quality of life among the pediatric survivors who undergo allo-HSCT. The significance of eosinophilia in peripheral blood after allo-HSCT remains to be elucidated. Previous studies found that eosinophilia after allo-HSCT reduced the risk of relapse of malignancies or non-relapse mortality and was also associated with cGVHD. However, the subjects of these reports were mainly adults and there were limited data focused on eosinophilia after allo-HSCT in children. The aim of this study was to evaluate whether an early elevation of absolute eosinophil count after allo-HSCT was associated with the development of cGVHD in children. METHODS A total of 165 consecutive patients who received allo-HSCT in our institute were included. Patients who had previously received allo-HSCT, relapsed or died before day 100, and did not achieve engraftment, were excluded. Eosinophilia was defined as an eosinophil count of more than 500 /µL in peripheral blood within 60 days of stem cell infusion. The primary outcome of the present study was the development of cGVHD. RESULTS In our cohort, eosinophilia was detected in 67 patients (40.6%), with a median maximum eosinophil count of 1,050 /µL (range, 504 - 7,812 /µL) at a median of 40 days (range, 24 - 59 days), while median maximum eosinophil count in those without eosinophilia was 189 /µL (range, 0 - 495 /µL). There were no significant differences between the patients with or without eosinophilia in disease type, graft source, donor type, conditioning regimen, and GVHD prophylaxis and treatment. The median time from day 0 to the date of diagnosis of cGVHD was 120 days (range, 100 - 1,302 days). The cumulative incidence rate of cGVHD at 3 years in the entire cohort was 39.3% (95% confidence interval: 32.2 - 47.4). Patients with eosinophilia had a higher 3-year cumulative incidence rate of cGVHD, compared with those without eosinophilia (49.1% vs. 32.7%, P = 0.007). Cox proportional hazards model analysis with adjustments for the potential confounding factors of HLA mismatched donor, unrelated donor, aGVHD, use of ATG as conditioning regimen revealed that eosinophilia was associated with an increased risk of cGVHD (adjusted hazard ratio: 2.12; 95% confidence interval: 1.16-3.85, P = 0.014). In the 165 patients who underwent allo-HSCT, there was no significant difference in 5-year overall survival (OS) between the eosinophilia and the non-eosinophilia groups (80.1% vs. 80.5%, P=0.76). In the 121 patients with malignant diseases, there was no significant difference in 5-year OS between the eosinophilia and the non-eosinophilia groups (75.5% vs. 76.3%, P = 0.70). There was also no significant difference in 5-year disease-free survival between the eosinophilia and the non-eosinophilia groups (69.1% vs. 77.1%, P = 0.34). On the other hand, there was a significant difference in 5-year GVHD-free relapse-free survival between the eosinophilia and the non-eosinophilia groups (41.1% vs. 62.4%, P = 0.04). DISCUSSION & CONCLUSION In the present study, we revealed that an early elevation of absolute eosinophil count after allo-HSCT was associated with cGVHD in children. Because eosinophilia after allo-HSCT occurs before the development of cGVHD, early eosinophilia after allo-HSCT may predict the development of cGVHD in children.