Association of Chronic Graft-versus-Host Disease with Late Effects following Allogeneic Hematopoietic Cell Transplantation for Children with Hematologic Malignancy

Chronic graft-versus-host disease (cGVHD) occurs in up to 25% of children following allogeneic hematopoietic cell transplantation (HCT) and continues to be a major cause of late morbidity and poor quality of life among long-term survivors of pediatric HCT. Late effects (LEs) of HCT are well documented in this population, and cGVHD has been identified as a risk factor for subsequent neoplasms (SNs) and several nonmalignant LEs (NM-LEs); however, the reported correlation between cGVHD and LEs varies among studies. We compared LEs occurring ≥2 years following childhood HCT for a hematologic malignancy in 2-year disease-free survivors with and without cGVHD and further evaluated the association of cGVHD features on the development of LEs. This systematic retrospective analysis used data from the Center of International Blood and Marrow Transplant Research (CIBMTR) on a large, representative cohort of 1260 survivors of pediatric HCT for hematologic malignancy to compare first malignant LEs and NM-LEs in those with a diagnosis of cGVHD and those who never developed cGVHD. The cumulative incidences of any first LE, SN, and NM-LE were estimated at 10 years after HCT, with death as a competing risk for patients with cGVHD versus no cGVHD. Cox proportional hazards models were used to evaluate the impact of cGVHD and its related characteristics on the development of first LEs. The estimated 10-year cumulative incidence of any LE in patients with and without cGVHD was 43% (95% CI, 38% to 48.2%) versus 32% (95% confidence interval [CI], 28.5% to 36.3%) (P.001), respectively. The development of cGVHD by 2 years post-HCT was independently associated with any LE (hazard ratio [HR], 1.38; 95% CI, 1.13 to 1.68; P = .001) and NM-LE (HR, 1.37; 95% CI, 1.10 to 1.70; P = .006), but not SN (HR, 1.30; 95% CI, .73 to 2.31; P = .38). cGVHD-related factors linked with the development of an NM-LE included having extensive grade cGVHD (HR, 1.60; 95% CI, 1.23 to 2.08; P = .0005), severe cGVHD (HR, 2.25; 95% CI, 1.60 to 3.17; P.0001), interrupted onset type (HR, 1.57; 95% CI, 1.21 to 2.05; P = .0008), and both mucocutaneous and visceral organ involvement (HR, 1.59; 95% CI, 1.24 to 2.03; P = .0002). No significant association between cGVHD-specific variables and SN was identified. Finally, the duration of cGVHD treatment of cGVHD with systemic immunosuppression was not significantly associated with SNs or NM-LEs. cGVHD was more closely associated with NM-LEs than with SNs among survivors of pediatric HCT for hematologic malignancy. In this analysis, the development of SNs was strongly associated with the use of myeloablative total body irradiation. cGVHD-related characteristics consistent with a state of greater immune dysregulation were more closely linked to NM-LEs.