Iron Overload Following Hematopoietic Stem Cell Transplantation: Prevalence, Severity, and Management in Children and Adolescents with Malignant and Nonmalignant Diseases

Abstract

Iron overload (IOL) is a frequently reported complication following hematopoietic stem cell transplantation (HSCT) that has been investigated extensively in the field of hemoglobinopathies but has not been thoroughly characterized after HSCT in pediatric malignancies. Our aim was to assess prevalence, severity, risk factors, and management of IOL, as defined using biochemical (serum ferritin) and radiologic tools (T2*-weighted magnetic resonance imaging [MRI]), in a cohort of pediatric patients who underwent HSCT for either malignant or benign diseases. This monocentric, retrospective, observational study included all the 163 patients alive and in continuous remission at 24 months post-HSCT out of the 219 consecutive children and adolescents who underwent HSCT at our institution between 2012 and 2018, were included in the study. IOL was classified into 4 categories: absent, mild, moderate, and severe. Among the 163 patients, 73% had some degree of IOL (mild in 37%, moderate in 29%, and severe in 7%). Moderate/severe IOL was more frequent among patients diagnosed with a malignant disease versus those with a benign disease (43% versus 19%; P = .0065). Trend lines for serum ferritin showed a "bell-shaped" distribution, with the highest levels recorded during the first 6 months post-HSCT, followed by a spontaneous reduction. Both pre-HSCT (1659 ng/mL versus 617 ng/mL; P < .001) and maximum post-HSCT (2473 ng/mL versus 1591 ng/mL; P < .001) median ferritin levels were statistically higher in the patients with malignancies. Radiologic assessment of IOL confirmed a more severe degree in patients with malignant disorders compared to those with benign disorders (median T2*-MRI, 4.20 msec [interquartile range (IQR), 3.0 to 6.40 msec] versus 7.40 msec [IQR, 4.90 to 11.00 msec]; P = .008). T2* levels were associated with the number of transfusions performed (P = .0006), with a steeper drop in T2* values for the first 20 transfusions and a milder slope for subsequent transfusions. T2* and ferritin values showed a statistically significant negative exponential relationship (P < .0001), although a ferritin level ≥1000 ng/mL showed poor specificity (48%) and low positive predictive value (53%) for discriminating moderate-to-severe IOL from absent-mild IOL as assessed by T2*-MRI, but with high sensitivity (92%) and negative predictive value (91%). In a multivariable model, >20 transfusions (odds ratio [OR], 4.07; 95% confidence interval [CI], 1.61 to 10.68; P = .003) and higher pre-HSCT ferritin level (P < .001) were associated with the risk of developing moderate-to-severe IOL. Use of a sibling donor (OR, .29; 95% CI, .10 to .77; P=.015) and a nonmalignancy (OR, .27; 95% CI, .08 to .82; P=.026) were protective factors. Phlebotomy (66%), low-dose oral chelators (9%), or a combined approach (25%) were started at a median of 12 months after HSCT in 78% of the patients with IOL. Six percent of the patients treated exclusively with phlebotomy (median, 14, significantly higher in patients >40 kg) discontinued phlebotomy owing to poor venous access, lack of compliance, or hypotension, whereas 39% of patients treated with chelators developed mild renal or hepatic side effects that resolved after tapering or discontinuation. Patients with malignancies showed statistically higher pre-HSCT and post-HSCT ferritin levels and lower T2* values. High ferritin level recorded on T2*-MRI showed unsatisfactory diagnostic accuracy in predicting IOL; thus, T2*-MRI should be considered a key tool for confirming IOL after HSCT in patients with an elevated serum ferritin level. IOL treatment is feasible after HSCT.