Eur J Pediatr DOI 10.1007/s00431-014-2295-5

ORIGINAL ARTICLE

An assessment of iron overload in children treated for cancer and nonmalignant hematologic disorders Jelena Rascon & Lina Rageliene & Sigita Stankeviciene & Darius Palionis & Algirdas Edvardas Tamosiunas & Nomeda Valeviciene & Tadas Zvirblis

Received: 29 November 2013 / Revised: 17 February 2014 / Accepted: 3 March 2014 # Springer-Verlag Berlin Heidelberg 2014

Abstract Our goal was to assess the natural fate of iron overload (IO) following transfusions of packed red blood cells (PRBCs) in children treated for cancer and nonmalignant disorders according to the intensity level of their treatment. Sixty-six children were followed up from February 2010 to March 2013. The transfusion burden was compared between three treatment intensity groups assigned according to the Intensity of Treatment Rating Scale 3.0 (ITR-3). IO was assessed by serial measurements of serum ferritin (SF) (n= Communicated by David Nadal J. Rascon (*) : L. Rageliene : S. Stankeviciene Center of Pediatric Oncology and Hematology, Children’s Hospital, Affiliate of Vilnius University Hospital Santariskiu Klinikos, Vilnius, Lithuania e-mail: [email protected] L. Rageliene e-mail: [email protected] S. Stankeviciene e-mail: [email protected]

66) and quantification of tissue iron by magnetic resonance imaging (MRI) (n=12). Of the children studied, 36 % (24/66) received moderately intensive treatment (level 2), 21 % (14/ 66) received very intensive treatment (level 3), and 42 % (28/ 66) received the most intensive treatment (level 4). The number of PRBC (p=0.016), the total transfused volume (p= 0.026), and transfused volume adjusted to body weight (p= 0.004) were significantly higher in the level 4 group. By the median follow-up time of 35.5 months (range 8–133), 21– 29 % of patients (including level 2 and level 3 children) had SF >1,000 μg/l 1 year after cessation of transfusions. The slowest decrease of SF was observed in the level 4 group. Initial MRI examination demonstrated either mild or moderate IO in the liver and spleen. Repetitive MRI showed significant improvement in relaxation time between the initial and follow-up MRI performances in the liver (5.9 vs. 8.6 ms, p= 0.03) and the spleen (4.3 vs. 8.8 ms, p=0.03). Conclusion: IO diminished over time, but in the level 4 patients, it was detectable for years after cessation of transfusions.

J. Rascon : L. Rageliene : D. Palionis : A. E. Tamosiunas : N. Valeviciene Faculty of Medicine, Vilnius University, Vilnius, Lithuania

Keywords Cancer survivors . Iron overload . Treatment intensity groups . Ferritin . MRI

D. Palionis e-mail: [email protected]

Abbreviations IO Iron overload MRI Magnetic resonance imaging PRBC Packed red blood cells SF Serum ferritin UIBC Unsaturated iron-binding capacity

A. E. Tamosiunas e-mail: [email protected] N. Valeviciene e-mail: [email protected] D. Palionis : A. E. Tamosiunas : N. Valeviciene Center of Radiology and Nuclear Medicine, Vilnius University Hospital Santariskiu Klinikos, Vilnius, Lithuania T. Zvirblis Hematology, Oncology and Transfusion Medicine Center, Vilnius University Hospital Santariskiu Klinikos, Vilnius, Lithuania e-mail: [email protected]

Introduction The current long-term survival rate of children suffering from malignant disorders is approximately 80 % [11]. Significant

Eur J Pediatr

progress in pediatric cancer survival rates is largely attributable to the intensification of chemotherapy and the appropriate use of hematopoietic stem cell transplantation (HSCT). An adequate transfusional support with packed red blood cells (PRBCs) is necessary in order to overcome anemic complications during chemotherapy and transplant procedure. In clinical practice, the transfusional burden is rarely being traced. However, in view of the increasing cure rates, an awareness of long-term sequelae of iatrogenic iron deposits in various organs and their potential impact on the quality of life is becoming more important. The adverse effect of iron overload (IO) caused by chronic transfusions has been well recognized in thalassemia [27, 6, 17] and myelodysplastic syndrome [20], but the number of studies exploring this issue in children treated for cancer or other nonmalignant disorders is still limited. Recently, Eng and Fish assessed the iron accumulation in children treated for acute lymphoblastic leukemia and demonstrated greater transfusion demand in high-risk patients compared to the standard risk [10]. Subsequent studies, which included a broad spectrum of pediatric malignancies, revealed a correlation between the intensity of treatment and the need for PRBC as well as the level of accumulated iron [25, 21]. A consensus opinion arose from the aforementioned analyses that there is a need to quantify the number of transfused PRBC and to monitor iatrogenic iron deposition. Studies that followed the accumulated IO over time found serum ferritin (SF) to remain elevated for years after the completion of chemotherapy [13, 12]. Along with the abnormal ferritin level, excessive iron deposition in the liver was observed by Halonen et al. who inspected liver biopsy specimens for intrahepatic iron [12]. Jastaniah et al. [13] used superconducting quantum interference devices (SQUID) for the same purposes. Both methods appeared to be reliable to prove iron accumulation in the liver. However, these techniques have limitations to quantifying intrahepatic iron routinely: a liver biopsy is an invasive procedure, while SQUID devices are available only in a very limited number of centers. Recently, magnetic resonance imaging (MRI) was demonstrated to be sensitive, noninvasive, and widely available to assess iron stores in the liver and myocardium [18, 23, 8]. In our study, we aimed to assess the natural fate of IO in children treated for cancer and nonmalignant disorders other than thalassemia according to the intensity level of treatment. Changes in the excess level of iron were assessed by serial measurements of serum iron parameters and parallel quantification of liver and cardiac iron deposits by means of MRI testing.

prospectively until March 1, 2013. The research project was conducted in accordance with Good Clinical Practice. The study design was approved by the Institutional Review Board and Ethics Committee. Written informed consent was obtained from parents or legal guardians of all the children prior to their enrollment in the study. During the study period, 66 children up to the age of 18, who had completed their treatment and were available for assessment on an outpatient basis, were recruited. At least one PRBC transfusion was required to be included in the study. Previous chelation therapy was assumed as exclusion criteria. No iron removal attempt was undertaken by patients during the study period. Demographic data (i.e., age, gender) and clinical characteristics (i.e., diagnosis, type of treatment, body weight, number of PRBC, the total transfused volume) as well as values of SF, serum iron, and unsaturated iron-binding capacity (UIBC) were retrieved from laboratory databases and patient records. A treatment intensity level was assigned according to the Intensity of Treatment Rating Scale version 3.0 (ITR-3) [14]. In order to monitor the kinetics of iron deposits, SF, serum iron, and UIBC were initially assessed at the time of recruitment. Subsequently, only ferritin was measured during regular patient visits. According to the study design, venous blood for research purposes was obtained only if the patient had central venous access. If not, the study variables could be measured only in case other clinically relevant analytics had to be obtained from the venous blood at the day of visit. SF level was measured using an “electrochemiluminescence immunoassay” (ECLIA) (Cobas E 411 Elecsys 1010; Roche Diagnostics GmbH, Manheim, Germany). The level of serum iron was analyzed by a colorimetric assay with FerroZine (Cobas Integra 700; Roche Diagnostics GmbH, Manheim, Germany). UIBC was analyzed using colorimetric assay direct determination with FerroZine (Cobas Integra 700; Roche Diagnostics GmbH, Manheim, Germany). The blood center of our institution manufactured leukodepleted PRBC using apheresis and an automated whole blood processing system pursuant to local requirements that defined target hematocrit level as 50–70 %. The total iron burden was calculated using the formulation described by Cohen et al. [7]: Iron transfused (mg)=Volume transfused× Hematocrit (Hct)×1.08, assuming mean empiric hematocrit to be 60 %. Projected transfused iron was adjusted to the recipient body weight measured at the time of the last PRBC transfusion. MRI methodology

Patients and methods The longitudinal observational study was initiated on February 1, 2010. The included patients were followed up

Patients were prospectively evaluated using a 1.5-T MRI (Avanto, Siemens) during 2011–2013. Tissue IO was assessed in a cohort of 12 randomly selected patients who could undergo MRI procedure without need for general anesthesia

Eur J Pediatr

(respiration and motion artifacts were not significant). Total MRI procedure time ranged from 20 to 30 min. Liver image acquisition for T2* evaluation was done with a single 10-mm slice through the central part of the liver (including both lobes) at 12 different echo times (TE 0.99– 16.5 ms) in transversal plane. Each image was acquired during an 11–13-s single breath hold using a dedicated gradient echo sequence (repetition time 200 ms, flip angle 20, matrix 128× 128 pixels, field of view 40 cm). The same sequence was used for the spleen in transversal plane. For evaluation of myocardial T2*, a short-axis mid-ventricular slice was acquired at eight separate echo times (TE 2.59–18.2 ms) with dedicated gradient echo sequence (flip angle 20, matrix 256×128 pixels, field of view 40 cm). T2* time was calculated from gradient echo images using CMRtools 2010 (©Cardiovascular Imaging Solutions). Cardiac T2* was evaluated from the intraventricular septum region (represents both ventricles) in short-axis view [5]. The left ventricle ejection fraction was evaluated from shortaxis images using Argus software (Siemens). Statistical analysis Descriptive statistics were used to describe demographic characteristics. A Mann-Whitney-Wilcoxon test was performed to evaluate differences between two independent groups. A comparison between more than two groups was performed using analysis of variance (ANOVA). Differences between independent two qualitative data groups were evaluated by means of chi-square test. Pearson correlation analysis was performed to evaluate relation between two variables. A two-sided p value less than 0.05 was considered to be significant. Statistical analysis was performed using the Statistical Analysis System (SAS) package version 9.2.

Results Sixty-six patients in total were evaluated. Within this group, 56 % represented different types of leukemia (37/66), 36 % of solid tumors (24/66), and 8 % (5/66) of nonmalignant disorders following HSCT (during the study period there was no transfusion-free patients with nonmalignant disorders without HSCT) (Table 1). Thirty-six percent of children (24/66) received moderately intensive treatment (level 2), 21 % (14/66) received very intensive treatment (level 3), and 42 % (28/66) received the most intensive treatment (level 4). Twenty-one patients (32 %) were grafted with either autologous or allogeneic hematopoietic stem cells. These children were assigned to the most intensive treatment level according to the definition [14]. There was no significant difference between treatment intensity groups in terms of the patients’ ages (p=0.058), the

time elapsed from the last PRBC transfusion, and the first SF measurement (p=0.333) and the follow-up period (p=0.275). An assessment of the transfusion burden revealed that the children treated in the most intensive way had the highest PRBC demand. Comparative analysis of the included variables between three treatment intensity groups (level 4 to level 3 to level 2, Table 1) showed that the number of transfused PRBC units (median 21.5 vs. 16 vs. 8, respectively, p=0.016), the total transfused volume (median 4,606.5 vs. 3,501 vs. 1,691, respectively, p=0.026), and the transfused PRBC volume adjusted to recipient body weight (median 159.6 vs. 117 vs. 87.3, respectively, p=0.004) were significantly superior for level 4. This was translated into the highest projected iron burden (median 3,233.8 vs. 2,453.7 vs. 1,187.1, respectively, p=0.026) and iron load adjusted for body weight (median 105.7 vs. 79.6 vs. 61.3, respectively, p=0.005). An initial evaluation of iron parameters at the time of enrollment into the study demonstrated SF to be highly elevated in all three groups. The ferritin level was significantly higher in the level 4 group as compared to the level 3 and level 2 groups (median 1,071.8 vs. 810 vs. 401.5, respectively, p= 0.008). Measurements of serum iron and UIBC at the same time point showed that both of the markers exceeded the normal range only in the level 4 group (Table 1). The difference between the three treatment intensity groups was not significant for either serum iron (p=0.363) or for UIBC (p= 0.402). Nevertheless, a higher number of PRBC transfusions correlated with a higher level of SF and serum iron (r=0.3944, p=0.0009, and r=0.3280, p=0.0244, respectively) and lower UIBC level (r=−0.3723, p=0.0009) (Fig. 1). Serial measurements of SF showed a gradual decrease of hyperferritinemia over time. Median ferritin value became lower than 1,000 μg/l approximately after 1 year of followup (Fig. 2). Twenty-nine percent of children (8/28) in the level 4 group had SF level over 1,000 μg/l following 1 year after cessation of PRBC transfusions as compared to 21 % (3/14) in the level 3 and 21 % (5/24) in the level 2 groups. After 2 years of follow-up, hyperferritinemia over 1,000 μg/l was found in 4 of 28 children (14 %) in the level 4 group and in 1 of 14 children (7 %) in the level 3 group. In one adolescent treated in the most intensive way sustained, hyperferritinemia >1,000 μg/l was still detectable 3 years after the last transfusion. Tissue iron evaluation using MRI Evaluation of tissue iron deposits using MRI T2* was performed in 12 patients. The median time from the last PRBC transfusion to the first MRI was 11.7 months (ranging from 1 to 44 months). In six children, an MRI scan was repeated approximately 1 year after the primary assessment with a median time of 13.5 months (ranging from 12 to 23 months). MRI evaluation was performed in older children. Compared

Eur J Pediatr Table 1 Clinical characteristics, transfusion burden, and iron status parameters Total (n=66)

Diagnoses ALLa (total) SR IR HR Relapse AML JMML Non-Hodgkin lymphoma Hodgkin lymphoma Wilms’ tumor Neuroblastoma Osteosarcoma Ewing sarcoma Rhabdomyosarcoma Adrenocortical carcinoma Germ cell tumor Nonmalignant disordersb

30 10 14 4 2 6 1 7 3 2 3 2 4

Level 2 (n=24)

Level 3 (n=14)

Level 4 (n=28)

17 10 7

8

5

7 1

6

1

1

1

3 2 6 1

3 2 4 1 1 1

5 22 11 11

25.7 02.–58.7

p value

3

1 1 1

HSCT (total) Autologous Allogeneic Age at diagnosis (years) Median 7.6 Range 0.1–16.8 Number of PRBC (units) Median 13 Range 1–144 Total volume of PRBC (ml) Median 2,404 Range 163–37,795 Transfused volume of PRBC per kg (ml/kg) Median 110.7 Range 4.1–899.9 Projected iron burden (mg) Median 1,687.3 Range 114.4–26,532.1 Adjusted iron burden per kg (mg/kg) Median 76.7 Range 2.9–631.7 Ferritin (normal range 4–327 μg/l) Median 701.2 Range 23.3–5,723.0 Serum iron (normal range 4.5–22.7 μmol/l) Median 19.3 Range 4.4–42.2 UIBC (normal range 20–66 μmol/l) Median Range

Treatment intensity level

5 21 11 11 4.0 0.7–15.0

11.5 2.6–15.2

7.9 0.1–16.8

0.058

8 1–22

16 2–48

21.5 1–144

0.016

1,691 163–5,449

3,501 720–11,827

4,606.5 260–37,795

0.026

87.3 4.1–181.3

117.0 42.4–394.2

159.6 7.4–899.9

0.004

1,187.1 114.4–3,825.2

2,457.7 505.4–8,302.6

3,233.8 182.5–26,532.1

0.026

61.3 2.9–127.2

79.6 31.6–276.8

105.7 5.2–631.7

0.005

401.5 31.2–2,404.0

810.0 98.0–5,723.0

1,071.8 23.3–5,657.1

0.008

17.8 4.4–37.7

16.0 11.1–36.0

24.2 5.2–42.2

0.363

30.4 3.7–58.7

26.7 3.4–35.1

15.2 0.2–53.6

0.402

Eur J Pediatr Table 1 (continued) Total (n=66)

Treatment intensity level Level 2 (n=24)

Time from the last PRBC transfusion until the first ferritin assessment (months) Median 9.6 13.9 Range 0.2–84.2 2.7–77.8 Follow-up (months) Median 35.5 37.5 Range 8–133 8–94

Level 3 (n=14)

Level 4 (n=28)

p value

9.95 0.2–27.9

3.2 0.5–84.2

0.333

35.5 15–60

36.5 8–133

0.275

a

Twenty-one from 30 patients with ALL were treated according to NOPHO ALL-2008 protocol, six children were treated according to BFM ALL-2000 protocol, one child had Ph-positive ALL, and one patient was treated according to NHL protocol because of mature B cell leukemia

b

Nonmalignant disorders consisted of SAA (n=2), DBA (n=1), Fanconi anemia (n=1), and MDS-RC (n=1). All five patients received allogeneic HSCT

to the remaining non-MRI cohort, the age at diagnosis of the MRI cohort was 10.3 versus 5.4 years, although the difference was not significant (Table 2). The median age at the first MRI examination was 12.9 years (ranging from 3.8 to 17 years). There was no significant difference in treatment intensity between MRI and non-MRI cohorts. However, transfusion intensity and subsequent iron burden differed significantly when comparing all included variables (Table 2). Clinically significant myocardial IO was excluded in all 12 patients validating T2* against cardiac function [3, 15]; in 11 patients, myocardial T2* was >20 ms (17–52.2) with an ejection fraction (EF) of >50 %, and one patient had mild myocardial iron overload (T2* value of 17 ms) with an EF of 54 %. Liver T2* values ranged from 1.9 to 22.9 ms, and according to Di Tucci et al. [9], two patients were suspected to have moderate (5–10 mg iron/g (dry liver)) and four patients mild (2– 5 mg iron/g (dry liver)) IO in the liver. According to Iron Health Alliance T2* validation versus biopsies [26, 29], three patients had moderate and six mild (due to upper T2* threshold at 11.4 ms) IO in the liver. Spleen initial T2* values ranged from 2.1 to 6.6 ms (no thresholds are available to determine IO degree). Preliminary data suggest that spleen iron deposition does not directly correlate to liver overload: five patients had lower initial spleen T2* and five higher compared to the liver. A follow-up MRI performed in six children revealed that IO decreased both in the liver and the spleen (Fig. 3a). Median relaxation time compared between the initial MRI and followup MRI improved significantly in the liver (5.9 vs. 8.6 ms, p= 0.03) and spleen (4.3 vs. 8.8 ms, p=0.03) (Fig. 3b). Myocardial tissue iron remained within normal range at initial and follow-up imaging (29.0 vs. 31.6 ms, p=0.81).

Discussion Our single-center study aimed to explore the natural fate of iatrogenic IO in children treated for the most common types of

pediatric cancer and nonmalignant disorders. The main point of interest was to evaluate whether the amount of accumulated iron differs depending on treatment intensity and what the natural course of iron deposits is. Sixty-six children followed up through the study period were available for the evaluation. Of note, no patients suffering from brain tumors (most of them assignable to the level 3 or 4) were represented. Likewise, no patients who received the least intensive treatment (level 1) were included in the study as none of them needed a PRBC transfusion. The largest group in our study was the one with the most intensive level of treatment (i.e., level 4). Level 4 accounted for 32 % (21/66) of the transplanted children. Along with cancer survivors, this group included the five recipients who underwent an allogeneic transplant for nonmalignant disorders. On the other hand, the abovementioned absence of neuro-oncology patients could also potentially contribute to the predominance of the most intensively treated children. As anticipated, the level 4 patients had the highest transfusion requirements. All evaluated variables of transfusion intensity and iron burden were significantly higher in the level 4 group (Table 1). Our findings are consistent with previously published data. Three retrospective studies reported that transfusion demand was associated with treatment intensity, and the highest projected iron burden was found in children treated in the most intensive way [21, 10, 25]. Evaluation of serum iron parameters showed significantly higher SF in the group receiving the most intensive treatment, though the values of serum iron and UIBC did not differ significantly between the three treatment groups. This could be explained by the methodology of quantification of serum iron and UIBC that is based on the measurement of transferrin-bound iron [24]. Some studies showed that UIBC is a sensitive marker for calculation of empty iron stores [4] and could be considered as an alternative to transferrin saturation [19]. We speculated that UIBC evaluation along with SF could be useful to prove hyperferritinemia due to IO, but not due to an underlying

Eur J Pediatr

Fig. 2 Kinetics of SF by serial assessment in the entire patient group (n= 66). One cross represents one serum sample. One patient had several SF measurements

Fig. 1 Relationship between iron parameters (a serum ferritin, b serum iron, c UIBC) and the number of transfused PRBC

inflammation or malignancy. Although we could not demonstrate the difference between serum iron and UIBC depending

on the treatment intensity level, a correlation between the number of transfused PRBC and all evaluated iron parameters was found (Fig. 1). Serial SF measurements over time showed that hyperferritinemia was decreasing gradually. The average level of ferritin fell below 1,000 μg/l approximately 1 year after the cessation of transfusions. Interestingly, we found that approximately one fifth of patients who received less intensive treatment (21 % of children in each level 2 and level 3 groups) was also heavily transfused and had severe hyperferritinemia detectable 1 year after cessation of transfusions. This finding suggests that the transfusion history should be carefully traced in all children to define a cohort for close monitoring of IO. In our study, the level 4 patients showed the slowest pattern of resolution of the accumulated iron; in one adolescent, an SF level over 1,000 μg/l was detected more than 3 years after the last PRBC transfusion. Slow resolution of hyperferritinemia could be related to the high initial SF (reflecting the degree of accumulated iron). A recent study by Amid et al. demonstrated that the age at diagnosis might affect the resolution of hyperferritinemia; greater change in body surface area was protective for the long-lasting hyperferritinemia [2]. In other words, smaller children utilize the accumulated iron for their needs to grow. In our study, there was no age difference between the three compared treatment groups, although some observations not reflected in this study also suggested that adolescents had a slower release of accumulated iron. The monitoring of tissue iron using MRI T2* showed a similar pattern of decrease of iatrogenic iron deposits expressed as a significant improvement in the liver and splenic relaxation time 1 year after the initial MRI T2* (Fig. 3). In our study, tissue iron evaluation using MRI T2* was performed only in older children who could undergo the

Eur J Pediatr Table 2 Characteristics of children who underwent tissue iron overload evaluation using MRI T2* technique (MRI cohort) and those who did not (non-MRI cohort) MRI cohort n=12 Treatment intensity, n (%) Level 2 3 (25) Level 3 1(8) Level 4 8 (67) Age at diagnosis (years) Median 10.3 Range 2.0–16.0 Number of PRBC (units)

Non-MRI cohort n=54

p value

21 (39) 13 (24) 20 (37)

0.157

5.4 0.1–16.8

0.1249

Median 28 12 Range 8–144 1–56 Total volume of PRBC (ml) Median 6,623 2,926 Range 1,984–37,795 163–13,696 Transfused volume of PRBC per kg (ml/kg) Median 200.6 106.1 Range 31.6–899.9 4.1–326.1 Projected iron burden (mg) Median 4,649 1,442.6 Range 1,392.8–26,532.1 114.4–9,614.6 Adjusted iron burden per kg (mg/kg) Median 142.5 74.2 Range 25.3–631.7 2.9–228.9 Ferritin (normal range 4–327 μg/l) Median 2,021.5 586.9 Range 285.0–5,723.0 23.3–5,657.1 Serum iron (normal range 4.5–22.7 μmol/l) Median 26.3 18.4 Range 21.1–42.2 4.4–34.1 UIBC (normal range 20–66 μmol/l) Median 6.9 28.5 Range 2.5–35.1 02.–58.7

0.0002

0.0004

0.0061

0.0004

0.0061

0.0006

0.0034

0.0014

procedure without general anesthesia. The major reason for this was our decision not to sedate small children solely for research purposes. This is the major limitation of our study as patient selection for MRI evaluation was biased with regard to age. The second limitation is the small number of children assessed by MRI, especially for the second time. Despite age differences (although not statistically significant), there was no difference in the treatment intensity level between MRI and non-MRI cohorts. Surprisingly, the MRI cohort was more intensively transfused and had a significantly greater iron burden and SF level (Table 2). The data published thus far have explored the influence of age at the time of diagnosis on IO only retrospectively and are in dispute. In the analysis of a mixed patient population, Ruccione et al. found that younger age at diagnosis was significant for projected iron burden [25].

Eng and Fish analyzed only ALL patients and reported no influence of age on the amount of blood received [10]. Amid et al. found that age, weight, and body surface area correlated with the level of SF at the end of treatment, though only the total transfused volume was independently significant for hyperferritinemia [2]. In addition, this study demonstrated that younger patients who had a more pronounced change in body surface area were less likely to have high SF after the completion of chemotherapy than older children. From clinical observations, it is well known that smaller children have a much higher tolerance to low hemoglobin levels and thus could require lesser transfusions than adolescents. Depending on the validation system, between six and nine children from the MRI cohort had measurements of the liver T2* that indicated mild or moderate IO. Liver MRI results can be compared to the liver biopsy. Three studies have confirmed that the use of a relaxometry liver MRI and biopsy values was strongly correlated [3, 1, 28]. Myocardial T2* values of 50 % for both patients. Data from β-thalassemia suggested that myocardial IO detectable by MRI was to be expected after at least 35 g of transfusional iron that corresponded to nearly 200 PRBC units [30]. The degree of liver IO and SF seem did not correlate with myocardial IO [30, 8]. A lower transfusion burden in our study population could justify normal myocardial T2* relaxation time in 11 of 12 patients. In contrast to our results, other groups did find excessive iron accumulation in cancer survivors. De Ville de Goyet et al. recently reported myocardial IO in 14 % of patients prospectively evaluated by MRI 1 year after treatment [8]. Interestingly, this study failed to find an association between myocardial T2* and transfused blood volume and SF, although such association could be proven for the liver T2*. Lutz et al. [18] also found myocardial IO in two of three cancer survivors who were heavily transfused. In our study, we could demonstrate iron accumulation in the spleen. Follow-up MRI examinations of both the liver and spleen after 1 year showed significant improvement in T2* relaxation indicating a reduction of iron stores (Fig. 3). However, sustained IO was still detectable in some patients. Recent studies evaluated spleen IO importance along with liver measurements. Kolnagou et al. [16] presented a comparison of iron load using MRI T2* and iron grading of stained biopsies, indicating that substantial but variable amounts of excessive iron are stored in the spleen (0–40 %) in addition to those in the liver. Spleen iron seems to be cleared faster than liver iron using effective chelation protocols. Normalization of the body iron stores at an early age could maintain the spleen in nearnormal capacity, and secondary effects such as cardiac and

Eur J Pediatr Fig. 3 Noninvasive evaluation of iron overload using MRI T2* technique in the liver, myocardium, and spleen. MRI 1 indicates initial MRI performed in 12 patients and MRI 2 repetitive imaging performed in six patients after 1 year. a Level of tissue iron deposits measured as relaxation time in 12 patients. b Difference in median relaxation time between MRI 1 and MRI 2 calculated for the six patients with paired MRI (Wilcoxon test)

other complications could be avoided. Oshtrakh and colleagues [22] reported preliminary data from a comparative study of human liver ferritin and spleen tissues from healthy individuals and a patient with primary myelofibrosis using Mössbauer spectroscopy. The results obtained demonstrated that the iron content in a patient’s spleen in the form of iron storage proteins was about ten times higher than that in normal tissue. They reported that IO in the case of patient with primary myelofibrosis may be related mainly to an increase in the ferritin content. Our initial data showed that spleen iron deposition does not correlate directly to liver overload, although more studies are needed to evaluate spleen IO correlation with ferritin and functional impact on the spleen. To our knowledge, no threshold data are available at the moment to determine normal and overloaded spleen parenchyma using MRI scanning. More comparable studies are needed to establish a reliable technique for IO diagnosis and its grade

evaluation as well as for the exact measurement of iron concentration. In summary, iatrogenic IO is most prominent in children who received the most intensive level of treatment (i.e., the level 4 patients). However, up to one fifth of patients treated less intensively is also exposed to numerous transfusions and excessive iron accumulation. Sustained IO detectable as persistent hyperferritinemia and intraorganic MRI changes decrease over time; though in heavily transfused children, it takes years until SF decreases below 1,000 μg/l. These findings are important for future studies that should clarify if a final clearance of the accumulated iron occurs or not. In case of irreversible IO, there is a need to assess the contribution of IO on late organ damage following cancer treatment. The long-term nature of iron deposits requires a special followup program of heavily transfused patients to prove the final clearance or persistence of IO.

Eur J Pediatr Acknowledgments The study was supported by a research grant from Novartis. The grant was used to purchase biochemical diagnostics. The financial support had no influence on study design, data collection, analysis, and interpretation on the writing of the report and on the decision to submit the report for publication. Conflict of interest The authors declare that the study does not contain any potential conflicts of interest and financial and personal relationships that might bias this work.

References 1. Alexopoulou E, Stripeli F, Baras P, Seimenis I, Kattamis A, Ladis V, Efstathopoulos E, Brountzos EN, Kelekis AD, Kelekis NL (2006) R2 relaxometry with MRI for the quantification of tissue iron overload in beta-thalassemic patients. J Magn Reson Imaging 23(2):163–170. doi:10.1002/jmri.20489 2. Amid A, Barrowman N, Vijenthira A, Lesser P, Mandel K, Ramphal R (2013) Risk factors for hyperferritinemia secondary to red blood cell transfusions in pediatric cancer patients. Pediatr Blood Cancer 60(10):1671–1675. doi:10.1002/pbc.24629 3. Anderson LJ, Holden S, Davis B, Prescott E, Charrier CC, Bunce NH, Firmin DN, Wonke B, Porter J, Walker JM, Pennell DJ (2001) Cardiovascular T2-star (T2*) magnetic resonance for the early diagnosis of myocardial iron overload. Eur Heart J 22(23): 2171–2179 4. Asberg A, Thorstensen K, Mikkelsen G, Asberg AE (2013) The diagnostic accuracy of unbound iron binding capacity (UIBC) as a test for empty iron stores. Scand J Clin Lab Invest 73(3):208–213. doi:10.3109/00365513.2013.765029 5. Carpenter JP, He T, Kirk P, Roughton M, Anderson LJ, de Noronha SV, Sheppard MN, Porter JB, Walker JM, Wood JC, Galanello R, Forni G, Catani G, Matta G, Fucharoen S, Fleming A, House MJ, Black G, Firmin DN, St Pierre TG, Pennell DJ (2011) On T2* magnetic resonance and cardiac iron. Circulation 123(14):1519– 1528. doi:10.1161/CIRCULATIONAHA.110.007641 6. Cianciulli P (2008) Treatment of iron overload in thalassemia. Pediatr Endocrinol Rev 6(Suppl 1):208–213 7. Cohen AR, Glimm E, Porter JB (2008) Effect of transfusional iron intake on response to chelation therapy in beta-thalassemia major. Blood 111(2):583–587. doi:10.1182/blood-2007-08-109306 8. de Ville de Goyet M, Moniotte S, Robert A, Dupont S, Vermylen C, Veyckemans F, Brichard B (2013) Iron overload in children undergoing cancer treatments. Pediatr Blood Cancer 60(12):1982–1987. doi:10.1002/pbc.24705 9. Di Tucci AA, Matta G, Deplano S, Gabbas A, Depau C, Derudas D, Caocci G, Agus A, Angelucci E (2008) Myocardial iron overload assessment by T2* magnetic resonance imaging in adult transfusion dependent patients with acquired anemias. Haematologica 93(9): 1385–1388. doi:10.3324/haematol.12759 10. Eng J, Fish JD (2011) Insidious iron burden in pediatric patients with acute lymphoblastic leukemia. Pediatr Blood Cancer 56(3):368–371. doi:10.1002/pbc.22851 11. Gatta G, Zigon G, Capocaccia R, Coebergh JW, Desandes E, Kaatsch P, Pastore G, Peris-Bonet R, Stiller CA (2009) Survival of European children and young adults with cancer diagnosed 1995-2002. Eur J Cancer 45(6):992–1005. doi:10.1016/j.ejca.2008.11.042 12. Halonen P, Mattila J, Suominen P, Ruuska T, Salo MK, Makipernaa A (2003) Iron overload in children who are treated for acute lymphoblastic leukemia estimated by liver siderosis and serum iron parameters. Pediatrics 111(1):91–96 13. Jastaniah W, Harmatz P, Pakbaz Z, Fischer R, Vichinsky E, Walters MC (2008) Transfusional iron burden and liver toxicity after bone

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24. 25.

26.

27.

marrow transplantation for acute myelogenous leukemia and hemoglobinopathies. Pediatr Blood Cancer 50(2):319–324. doi:10.1002/ pbc.21260 Kazak AE, Hocking MC, Ittenbach RF, Meadows AT, Hobbie W, DeRosa BW, Leahey A, Kersun L, Reilly A (2012) A revision of the intensity of treatment rating scale: classifying the intensity of pediatric cancer treatment. Pediatr Blood Cancer 59(1):96–99. doi:10.1002/ pbc.23320 Kirk P, Roughton M, Porter JB, Walker JM, Tanner MA, Patel J, Wu D, Taylor J, Westwood MA, Anderson LJ, Pennell DJ (2009) Cardiac T2* magnetic resonance for prediction of cardiac complications in thalassemia major. Circulation 120(20):1961–1968. doi:10.1161/ CIRCULATIONAHA.109.874487 Kolnagou A, Michaelides Y, Kontoghiorghe CN, Kontoghiorghes GJ (2013) The importance of spleen, spleen iron, and splenectomy for determining total body iron load, ferrikinetics, and iron toxicity in thalassemia major patients. Toxicol Mech Methods 23(1):34–41. doi: 10.3109/15376516.2012.735278 Kwiatkowski JL, Kim HY, Thompson AA, Quinn CT, Mueller BU, Odame I, Giardina PJ, Vichinsky EP, Boudreaux JM, Cohen AR, Porter JB, Coates T, Olivieri NF, Neufeld EJ (2012) Chelation use and iron burden in North American and British thalassemia patients: a report from the Thalassemia Longitudinal Cohort. Blood 119(12): 2746–2753. doi:10.1182/blood-2011-04-344507 Lutz K, von Komorowski G, Durken M, Engelhardt R, Dinter DJ (2008) Myocardial iron overload in transfusion-dependent pediatric patients with acute leukemia. Pediatr Blood Cancer 51(5):691–693. doi:10.1002/pbc.21663 Murtagh LJ, Whiley M, Wilson S, Tran H, Bassett ML (2002) Unsaturated iron binding capacity and transferrin saturation are equally reliable in detection of HFE hemochromatosis. Am J Gastroenterol 97(8):2093–2099. doi:10.1111/j.1572-0241.2002. 05927.x Neukirchen J, Fox F, Kundgen A, Nachtkamp K, Strupp C, Haas R, Germing U, Gattermann N (2012) Improved survival in MDS patients receiving iron chelation therapy—a matched pair analysis of 188 patients from the Dusseldorf MDS registry. Leuk Res 36(8): 1067–1070. doi:10.1016/j.leukres.2012.04.006 Nottage K, Gurney JG, Smeltzer M, Castellanos M, Hudson MM, Hankins JS (2013) Trends in transfusion burden among long-term survivors of childhood hematological malignancies. Leuk Lymphoma 54(8):1719–1723. doi:10.3109/10428194.2012.750724 Oshtrakh MI, Alenkina IV, Vinogradov AV, Konstantinova TS, Kuzmann E, Semionkin VA (2013) Mossbauer spectroscopy of the iron cores in human liver ferritin, ferritin in normal human spleen and ferritin in spleen from patient with primary myelofibrosis: preliminary results of comparative analysis. Biometals 26(2):229–239. doi: 10.1007/s10534-012-9602-2 Reddy KS, Eng J, Carlson C, Ginsberg J, Fish J (2011) Impact of transfusional support for childhood cancer—peering behind the iron curtain. Pediatr Blood Cancer 56:908 Roche DIA customer self-service portal. https://cssportal.roche.com/ eld/lt/lt/DocumentResult/ResultList. Accessed 10 Aug 2013 Ruccione KS, Mudambi K, Sposto R, Fridey J, Ghazarossian S, Freyer DR (2012) Association of projected transfusional iron burden with treatment intensity in childhood cancer survivors. Pediatr Blood Cancer 59(4):697–702. doi:10.1002/pbc.24046 St Pierre TG, Clark PR, Chua-anusorn W, Fleming AJ, Jeffrey GP, Olynyk JK, Pootrakul P, Robins E, Lindeman R (2005) Noninvasive measurement and imaging of liver iron concentrations using proton magnetic resonance. Blood 105(2):855–861. doi:10.1182/blood2004-01-0177 Taher AT, Musallam KM, Karimi M, Cappellini MD (2012) Contemporary approaches to treatment of beta-thalassemia intermedia. Blood Rev 26(Suppl 1):S24–27. doi:10.1016/S0268960X(12)70008-5

Eur J Pediatr 28. Voskaridou E, Douskou M, Terpos E, Papassotiriou I, Stamoulakatou A, Ourailidis A, Loutradi A, Loukopoulos D (2004) Magnetic resonance imaging in the evaluation of iron overload in patients with beta thalassaemia and sickle cell disease. Br J Haematol 126(5):736–742. doi:10.1111/j.1365-2141.2004.05104.x 29. Wood JC, Enriquez C, Ghugre N, Tyzka JM, Carson S, Nelson MD, Coates TD (2005) MRI R2 and R2* mapping accurately estimates

hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients. Blood 106(4):1460–1465. doi:10.1182/ blood-2004-10-3982 30. Wood JC, Origa R, Agus A, Matta G, Coates TD, Galanello R (2008) Onset of cardiac iron loading in pediatric patients with thalassemia major. Haematologica 93(6):917–920. doi:10.3324/ haematol.12513