- Email: [email protected]
Longitudinal analysis of heart and liver iron in thalassemia major patients according to chelation treatment
Blood Cells, Molecules and Diseases 51 (2013) 142–145
Contents lists available at ScienceDirect
Blood Cells, Molecules and Diseases journal homepage: www.elsevier.com/locate/bcmd
Longitudinal analysis of heart and liver iron in thalassemia major patients according to chelation treatment F. Danjou a,⁎, R. Origa b, F. Anni a, L. Saba c, S. Cossa d, G. Podda d, R. Galanello a,b,c a Dipartimento di Sanità Pubblica, Medicina Clinica e Molecolare, Sezione di Scienze Biomediche e Biotecnologie, Università di Cagliari, Ospedale Regionale per le Microcitemie ASL8, Cagliari, Italy b Clinica Pediatrica 2a, Ospedale Regionale per le Microcitemie ASL8, Cagliari, Italy c Ambulatorio di Ematologia Pediatrica, Ospedale Regionale per le Microcitemie ASL8, Cagliari, Italy d Struttura Complessa Radiologia-Presidio Ospedaliero “Businco”, Cagliari, Italy
a r t i c l e
i n f o
Article history: Submitted 19 March 2013 Revised 7 May 2013 Available online 28 June 2013 (Communicated by Nancy Berliner, MD, 29 May 2013) Keywords: Iron chelation Beta-thalassemia Heart siderosis Liver iron concentration Magnetic resonance
a b s t r a c t Iron chelators and nuclear magnetic resonance imaging (MRI) techniques for assessing iron loading in liver and heart have greatly improved survival of thalassemic patients suffering iron overload-associated cardiomyopathy. However, the correlation between liver iron concentration and myocardial siderosis is ambiguous. Using an objective metric of time delay, scientists have demonstrated a lag in the loading and unloading of cardiac iron with respect to that of the liver. In the present study, we further tested this hypothesis with different chelation treatments. We analyzed the effect of three chelating treatment approaches on liver and cardiac iron content in 24 highly compliant patients who underwent 3 or more MRIs under each chelation treatment. Of the 84 MRIs considered, 32 were performed on deferoxamine (DFO — 8 patients), 24 on deferiprone (DFP — 7 patients), and 28 on combined therapy (DFO + DFP — 9 patients). In patients treated with DFO, changes in cardiac iron significantly lagged changes in liver iron but the opposite pattern was observed in patients treated with DFP (p = 0.005), while combined therapy showed a pattern in-between DFO and DFP. We conclude that the temporality of changes of cardiac and liver iron is chelator-dependent, so that chelation therapy can be tailored to balance iron elimination from the liver and the heart. © 2013 Elsevier Inc. All rights reserved.
Introduction Iron-induced cardiomyopathy is the single most frequent cause of premature death in patients with transfusion-dependent thalassemia [1]. The recent development of magnetic resonance imaging (MRI) techniques for assessing iron concentration in the heart provides a tool for early diagnosis of cardiac iron loading and for the monitoring of cardiac iron chelation efficacy [2]. Cross-sectional analyses of MRI assessments of cardiac and liver iron concentration (LIC) have shown a poor correlation between LIC and myocardial siderosis with some patients developing cardiac deposition in the absence of significant hepatic iron overload [3]. On the other hand, using an objective metric of time delay, Noetzli et al. have demonstrated a lag in the loading and unloading of cardiac iron with respect to liver iron in most of their cohort of 38 patients [4]. However, the type and pattern of chelation was not considered in the latter study. We therefore tested the hypothesis that cardiac iron removal effectively lags liver iron removal when one considers different chelation treatments. Precise knowledge of accumulation and clearance differential rates comparing heart and ⁎ Corresponding author at: Ospedale Regionale Microcitemie, Via Jenner s/n, 09121 Cagliari, Italy. Fax: +39 070 6095509. E-mail address: [email protected] (F. Danjou). 1079-9796/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.bcmd.2013.06.001
liver could help understand and manage iron chelation treatment in thalassemia patients. Materials and methods Forty-one patients with transfusion-dependent thalassemia followed at the Thalassemia Unit, Ospedale Regionale per le Microcitemie, Cagliari, Italy, underwent 3 or more MRI assessments of cardiac and liver iron content while receiving the same chelation treatment between 2003 and 2010. Twenty-four of them were regularly chelated with a treatment compliance greater than 70% during the period of this study. Each one underwent all MRIs with the same chelating treatment. Compliance was calculated upon the annual drug prescribed to dispensed ratio registered by the hospital pharmacy. These patients underwent a total of 84 MRIs: 28 while on combined therapy (8 patients), 32 while on DFO (9 patients) and 24 while on DFP (7 patients). Demographics of these patients are summarized in Table 1. The average time between a patient's first and last MRI was 4.2 ± 1.3 years (1.6–6.4);each MRI was done at an interval ranging from 3.5 to 56.2 months (median = 18 months). The interval was not significantly dissimilar among groups of chelators (medians: DFO = 16 months, DFO + DFP = 15 months, DFP = 20 months). Fifty percent of patients underwent three MRIs, 41.7% underwent four and 8.3% underwent five.
F. Danjou et al. / Blood Cells, Molecules and Diseases 51 (2013) 142–145
143
Table 1 Population characteristics.
Total Treatment a
DFO DFO + DFP DFP
Age (years)b Mean ± SD
Cardiac R2 (Hz)b Mean ± SD
LIC (mg/g dry weight)b Mean ± SD
Serum ferritin (ng/ml)b Median [25th%–75th%]
Annual blood consumption (ml/kg)c Mean ± Std. Dev.
27.7 26.0 29.8 27.5
94.8 103.0 107.1 70.2
4.9 5.6 4.8 3.9
1465 1720 660 1510
148.1 146.5 141.4 158.7
± ± ± ±
4.7 5.4 3.7 4.6
± ± ± ±
42.6 39.6 49.6 31.1
± ± ± ±
3.0 3.0 3.9 1.4
[630–2405] [1170–4600] [305–2215] [920–2160]
± ± ± ±
28.9 35.2 23.0 24.1
Annual blood consumption: volume of packed red cells (ml blood × hematocrit %) transfused in a year per kilogram of weight. a Only one significant difference was found, for annual blood consumption between DFO + DFP and DFP treatments (Mann–Whitney U-test). b At first MRI. c During study.
LIC was estimated using liver T2* measurements, and converted to mg/g dry weight using the following formula: LIC = 0.202 + 25.4/ liver T2* [5]. Cardiac T2* was determined by previously validated methods [6] and linearized as R2* (=1000/T2*) for use in calculations. The metric used to detect temporality of changes in cardiac and liver iron levels was calculated using the method described by Noetzli et al. [4], using the percentage variation of LIC and R2* to standardize metrics (see supplementary methods). Applying this approach, as explained in the original article, the area under the curve (AUC) calculated for each patient reflects the delay between a change in the two measurements. A zero total area corresponds to the lack of delay, whereas if heart lags liver, the AUC has a positive sign; and, if liver lags heart, the AUC has a negative sign. The proportion of positive AUCs and their distribution symmetry were compared using Pearson's Chi2 test and the Mann–Whitney U-test. Potential predictors of AUCs positivity were investigated with a binary logistic regression model. Goodness of fit of the model was assessed through Hosmer & Lemeshow test whereas Nagelkerke R2 was used to measure how useful explanatory variables were in predicting the outcome. All statistical procedures were performed using SPSS version 17.0 software package (SPSS Inc., Chicago, Illinois, USA).
have a significant effect on AUC positivity with an O.R. = 20.0 for DFP versus DFO treatment (p = 0.027, Nagelkerke R2 = 0.324). Knowledge of the chelator used is, therefore, essential to describe and compare the speed of decreases in iron loads between heart and liver, as much of this process likely reflects chelating treatment specificities. Finally, to understand if these different temporalities between treatments involved mainly loading or unloading of iron in the small sample under study, we plotted median cardiac R2 in function of median LIC for each treatment group for the first three MRIs (available for all patients in each group and representing 86% of all MRIs). Results are presented in Fig. 2 and show cardiac unloading prevalence in the DFP and combined therapy, while DFO demonstrates a prevalent cardiac iron loading process, whereas all treatments exhibit a prevalence of liver loading process. However the only significant variations were observed between second and third MRI for DFO + DFP on cardiac R2 (p = 0.050; Wilcoxon signed-rank test) and between first and second MRI for DFP on LIC (p = 0.018; Wilcoxon signed-rank test). These data complement previous results by showing that in DFP group liver iron lags cardiac iron in a prevalent unloading process of the heart, showing a pattern of cardiac specificity for DFP, as opposed to DFO. Furthermore, combined therapy demonstrates a pattern of cardiac unloading similar to DFP and a pattern of liver loading similar to DFO.
Results
Discussion
Among the AUCs from the 24 time courses of the patients considered in the study, 9 (37.5%) had an overall negative area (change in liver iron lagging a change in cardiac iron) and 15 (62.5%) had an overall positive area (change in cardiac iron lagging a change in liver iron) (p = 0.221, Pearson's Chi2 test). In this sample, distribution was positively skewed (skewness = 3.3 [Std. Err. = 0.47]; mean = 85.9; median = 25.3) and two large areas were observed in positive direction while AUCs were otherwise similar in both directions. The magnitude of the positive areas averaged 181.7 (median = 87.2; minimum = 12.2; maximum = 1202.5), and the magnitude of the negative areas averaged −73.8 (median = −90.4; minimum = −141.5; maximum = −3.3) resulting in a significant difference between positive and negative ranked AUCs (p b 0.001 Mann–Whitney U-test). Therefore, as in the previously reported work, overall AUCs data (i.e. both their sign, which represent the time delay between organs and their magnitude, which represent how much one organ lags the other) showed that a change in cardiac iron content significantly lags liver iron content [4]. To take in consideration the influence of the chelating treatment on the kinetics of iron loading and unloading in heart and liver, we further analyzed AUCs distribution for each of the three different chelators (Fig. 1). From such analysis, we observed that AUCs clearly behave differently upon treatment (all two-by-two comparisons and single group versus null variations were tested, see Table 2 for details). This finding challenged the first observation on overall AUCs data. To further assess the role of treatments, we developed a binary logistic model for AUC positivity including treatment, gender, time of study, age, cardiac R2* and LIC at first MRI. Treatment was the only variable to
In the present study we tested the hypothesis that, in regularly transfused thalassemia patients, cardiac iron loading or unloading lags
Fig. 1. Box-and-whisker plots of AUCs for each treatment group: central band characterizes the median, box boundaries delimit interquartile range (IQR), whiskers represent 1.5 IQR, whereas outliers are symbolized with circles and extremes (>3 IQRs) with a star. p-value is referred to Mann–Whitney U-test (p = 0.013 not considering the two outliers under DFO treatment).
144
F. Danjou et al. / Blood Cells, Molecules and Diseases 51 (2013) 142–145
Table 2 AUCs magnitude and signs according to chelation treatment. Median
Max.
Min.
AUC differences
Proportion of positive and negative AUCs (Exact test)
Mann–Whitney U-test DFO DFO + DFP DFP
−97.96 DFO vs. DFO + DFP: p = 0.102 115.68 −125.36 DFO + DFP vs. DFP: p = 0.083 21.46 −141.54 –
101.94 1202.45 41.67 −31.42
Wilcoxon signed rank test DFO vs. DFP: p = 0.005 – –
liver iron as formulated by Noetzli et al. [4] In our patient population as a whole, we confirmed a delay in cardiac iron changes. However, the organ-specific mechanisms of iron intake and iron elimination suggested by Noetzli et al. are only one of the key factors able to explain this observation. Although the possible role of genetic factors in organ deposition has been studied [7], in chelated patients, iron overload is usually inversely related to the compliance with chelation. To test the influence of the chelating treatments we selected 24 patients with good compliance with chelation treatment who underwent at least 3 MRIs under the same chelation treatment between 2003 and 2010, and analyzed their heart and liver iron concentration over time. We found that temporality of liver and cardiac iron changes significantly differ between DFO and DFP treatments. In patients treated with DFO, three quarters of AUCs were positive, meaning that cardiac iron significantly lagged liver iron changes, as demonstrated by Noetzli et al. [4] Such result could partly explain why many patients treated with DFO have heart iron overload in absence of relevant liver siderosis [8]. The group of patients treated with DFP showed an opposite pattern, with three quarters of negative AUCs, meaning that liver iron significantly lagged cardiac iron. This is most likely due to a rapid and selective cardiac iron removal in the DFP group, as demonstrated in previous randomized studies comparing the effects of DFP and DFO on heart and liver MRI [9]. These results are further supported by the fact that patients under combined therapy demonstrated, as expected, a pattern that combines both DFO and DFP characteristics (both for AUCs analysis and in term of organs unloading events prevalence). In fact, the association between DFO and DFP has been found able to remove both liver and heart iron in
DFO vs. 0: p = 0.028 DFO + DFP vs. 0: p = 0.401 DFP vs. 0: p = 0.063
DFO vs. DFO + DFP: p = 0.294 DFO + DFP vs. DFP: p = 0.315 –
DFO vs. DFP: p = 0.035 – –
DFO vs. expected: p = 0.039 DFO + DFP vs. expected: p = 0.727 DFP vs. expected: p = 0.453
an additive or synergistic way and more rapidly than the monotherapies [9–11]. It is well-known that body iron is made of diverse pools that are differently regulated, and several studies have already shown that iron chelators do not have the same effect on the different pools of iron [12]. The present data are coherent with such findings showing that DFO and DFP target different iron pools, with DFP targeting more selectively intracellular labile iron responsible of cardiac toxicity and combined therapy affecting different pools of iron at the same time [13–15]. In conclusion, our study clearly shows that the temporality of changes between cardiac and liver iron reflects chelator's access to labile cardiac iron, instead of unrevealing a physiological mechanism. To determine the existence of a physiological temporality between cardiac and liver iron changes, it seems therefore essential to, ideally, study a cohort of non-chelated subjects.
Conflict of interest Authors declare no conflict of interest.
Acknowledgments This research is in part supported by grant: LR.7 Agosto 2007, N. 7, Promozione della Ricerca Scientifica e dell'innovazione Tecnologica in Sardegna, from the Regione Autonoma della Sardegna L.R. 11 1990 and by Telethon (Project number GGP08221).
Fig. 2. The small diamonds, squares and triangles represent the median cardiac R2. The median cardiac R2 is shown as a function of the median liver iron content for each treatment group based on the first three MRIs.
F. Danjou et al. / Blood Cells, Molecules and Diseases 51 (2013) 142–145
Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.bcmd.2013.06.001.
References [1] B. Modell, M. Khan, M. Darlison, et al., Improved survival of thalassaemia major in the UK and relation to t2* cardiovascular magnetic resonance, J. Cardiovasc. Magn. Reson. 10 (2008) 42. [2] D.J. Pennell, T2* magnetic resonance and myocardial iron in thalassemia, Ann. N. Y. Acad. Sci. 1054 (2005) 373–378. [3] L.J. Anderson, S. Holden, B. Davis, et al., Cardiovascular t2-star (t2*) magnetic resonance for the early diagnosis of myocardial iron overload, Eur. Heart J. 22 (2001) 2171–2179. [4] L.J. Noetzli, S.M. Carson, A.S. Nord, T.D. Coates, J.C. Wood, Longitudinal analysis of heart and liver iron in thalassemia major, Blood 112 (2008) 2973–2978. [5] J.C. Wood, C. Enriquez, N. Ghugre, et al., Mri r2 and r2* mapping accurately estimates hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients, Blood 106 (2005) 1460–1465. [6] M.A. Tanner, T. He, M.A. Westwood, D.N. Firmin, D.J. Pennell, Multi-center validation of the transferability of the magnetic resonance t2* technique for the quantification of tissue iron, Haematologica 91 (2006) 1388–1391.
145
[7] R. Origa, S. Satta, G. Matta, R. Galanello, Glutathione s-transferase gene polymorphism and cardiac iron overload in thalassaemia major, Br. J. Haematol. 142 (2008) 143–145. [8] M.A. Tanner, R. Galanello, C. Dessi, et al., Myocardial iron loading in patients with thalassemia major on deferoxamine chelation, J. Cardiovasc. Magn. Reson. 8 (2006) 543–547. [9] M.A. Tanner, R. Galanello, C. Dessi, et al., A randomized, placebo-controlled, double-blind trial of the effect of combined therapy with deferoxamine and deferiprone on myocardial iron in thalassemia major using cardiovascular magnetic resonance, Circulation 115 (2007) 1876–1884. [10] R. Galanello, A. Agus, S. Campus, et al., Combined iron chelation therapy, Ann. N. Y. Acad. Sci. 1202 (2010) 79–86. [11] P. Evans, R. Kayyali, R.C. Hider, J. Eccleston, J.B. Porter, Mechanisms for the shuttling of plasma non-transferrin-bound iron (NTBI) onto deferoxamine by deferiprone, Transl. Res. 156 (2010) 55–67. [12] J.B. Porter, Optimizing iron chelation strategies in beta-thalassaemia major, Blood Rev. 23 (Suppl. 1) (2009) S3–S7. [13] L.J. Anderson, B. Wonke, E. Prescott, et al., Comparison of effects of oral deferiprone and subcutaneous desferrioxamine on myocardial iron concentrations and ventricular function in beta-thalassaemia, Lancet 360 (2002) 516–520. [14] C. Hershko, G. Link, A.M. Konijn, Z.I. Cabantchik, Objectives and mechanism of iron chelation therapy, Ann. N. Y. Acad. Sci. 1054 (2005) 124–135. [15] D.J. Pennell, V. Berdoukas, M. Karagiorga, et al., Randomized controlled trial of deferiprone or deferoxamine in beta-thalassemia major patients with asymptomatic myocardial siderosis, Blood 107 (2006) 3738–3744.
Contents lists available at ScienceDirect
Blood Cells, Molecules and Diseases journal homepage: www.elsevier.com/locate/bcmd
Longitudinal analysis of heart and liver iron in thalassemia major patients according to chelation treatment F. Danjou a,⁎, R. Origa b, F. Anni a, L. Saba c, S. Cossa d, G. Podda d, R. Galanello a,b,c a Dipartimento di Sanità Pubblica, Medicina Clinica e Molecolare, Sezione di Scienze Biomediche e Biotecnologie, Università di Cagliari, Ospedale Regionale per le Microcitemie ASL8, Cagliari, Italy b Clinica Pediatrica 2a, Ospedale Regionale per le Microcitemie ASL8, Cagliari, Italy c Ambulatorio di Ematologia Pediatrica, Ospedale Regionale per le Microcitemie ASL8, Cagliari, Italy d Struttura Complessa Radiologia-Presidio Ospedaliero “Businco”, Cagliari, Italy
a r t i c l e
i n f o
Article history: Submitted 19 March 2013 Revised 7 May 2013 Available online 28 June 2013 (Communicated by Nancy Berliner, MD, 29 May 2013) Keywords: Iron chelation Beta-thalassemia Heart siderosis Liver iron concentration Magnetic resonance
a b s t r a c t Iron chelators and nuclear magnetic resonance imaging (MRI) techniques for assessing iron loading in liver and heart have greatly improved survival of thalassemic patients suffering iron overload-associated cardiomyopathy. However, the correlation between liver iron concentration and myocardial siderosis is ambiguous. Using an objective metric of time delay, scientists have demonstrated a lag in the loading and unloading of cardiac iron with respect to that of the liver. In the present study, we further tested this hypothesis with different chelation treatments. We analyzed the effect of three chelating treatment approaches on liver and cardiac iron content in 24 highly compliant patients who underwent 3 or more MRIs under each chelation treatment. Of the 84 MRIs considered, 32 were performed on deferoxamine (DFO — 8 patients), 24 on deferiprone (DFP — 7 patients), and 28 on combined therapy (DFO + DFP — 9 patients). In patients treated with DFO, changes in cardiac iron significantly lagged changes in liver iron but the opposite pattern was observed in patients treated with DFP (p = 0.005), while combined therapy showed a pattern in-between DFO and DFP. We conclude that the temporality of changes of cardiac and liver iron is chelator-dependent, so that chelation therapy can be tailored to balance iron elimination from the liver and the heart. © 2013 Elsevier Inc. All rights reserved.
Introduction Iron-induced cardiomyopathy is the single most frequent cause of premature death in patients with transfusion-dependent thalassemia [1]. The recent development of magnetic resonance imaging (MRI) techniques for assessing iron concentration in the heart provides a tool for early diagnosis of cardiac iron loading and for the monitoring of cardiac iron chelation efficacy [2]. Cross-sectional analyses of MRI assessments of cardiac and liver iron concentration (LIC) have shown a poor correlation between LIC and myocardial siderosis with some patients developing cardiac deposition in the absence of significant hepatic iron overload [3]. On the other hand, using an objective metric of time delay, Noetzli et al. have demonstrated a lag in the loading and unloading of cardiac iron with respect to liver iron in most of their cohort of 38 patients [4]. However, the type and pattern of chelation was not considered in the latter study. We therefore tested the hypothesis that cardiac iron removal effectively lags liver iron removal when one considers different chelation treatments. Precise knowledge of accumulation and clearance differential rates comparing heart and ⁎ Corresponding author at: Ospedale Regionale Microcitemie, Via Jenner s/n, 09121 Cagliari, Italy. Fax: +39 070 6095509. E-mail address: [email protected] (F. Danjou). 1079-9796/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.bcmd.2013.06.001
liver could help understand and manage iron chelation treatment in thalassemia patients. Materials and methods Forty-one patients with transfusion-dependent thalassemia followed at the Thalassemia Unit, Ospedale Regionale per le Microcitemie, Cagliari, Italy, underwent 3 or more MRI assessments of cardiac and liver iron content while receiving the same chelation treatment between 2003 and 2010. Twenty-four of them were regularly chelated with a treatment compliance greater than 70% during the period of this study. Each one underwent all MRIs with the same chelating treatment. Compliance was calculated upon the annual drug prescribed to dispensed ratio registered by the hospital pharmacy. These patients underwent a total of 84 MRIs: 28 while on combined therapy (8 patients), 32 while on DFO (9 patients) and 24 while on DFP (7 patients). Demographics of these patients are summarized in Table 1. The average time between a patient's first and last MRI was 4.2 ± 1.3 years (1.6–6.4);each MRI was done at an interval ranging from 3.5 to 56.2 months (median = 18 months). The interval was not significantly dissimilar among groups of chelators (medians: DFO = 16 months, DFO + DFP = 15 months, DFP = 20 months). Fifty percent of patients underwent three MRIs, 41.7% underwent four and 8.3% underwent five.
F. Danjou et al. / Blood Cells, Molecules and Diseases 51 (2013) 142–145
143
Table 1 Population characteristics.
Total Treatment a
DFO DFO + DFP DFP
Age (years)b Mean ± SD
Cardiac R2 (Hz)b Mean ± SD
LIC (mg/g dry weight)b Mean ± SD
Serum ferritin (ng/ml)b Median [25th%–75th%]
Annual blood consumption (ml/kg)c Mean ± Std. Dev.
27.7 26.0 29.8 27.5
94.8 103.0 107.1 70.2
4.9 5.6 4.8 3.9
1465 1720 660 1510
148.1 146.5 141.4 158.7
± ± ± ±
4.7 5.4 3.7 4.6
± ± ± ±
42.6 39.6 49.6 31.1
± ± ± ±
3.0 3.0 3.9 1.4
[630–2405] [1170–4600] [305–2215] [920–2160]
± ± ± ±
28.9 35.2 23.0 24.1
Annual blood consumption: volume of packed red cells (ml blood × hematocrit %) transfused in a year per kilogram of weight. a Only one significant difference was found, for annual blood consumption between DFO + DFP and DFP treatments (Mann–Whitney U-test). b At first MRI. c During study.
LIC was estimated using liver T2* measurements, and converted to mg/g dry weight using the following formula: LIC = 0.202 + 25.4/ liver T2* [5]. Cardiac T2* was determined by previously validated methods [6] and linearized as R2* (=1000/T2*) for use in calculations. The metric used to detect temporality of changes in cardiac and liver iron levels was calculated using the method described by Noetzli et al. [4], using the percentage variation of LIC and R2* to standardize metrics (see supplementary methods). Applying this approach, as explained in the original article, the area under the curve (AUC) calculated for each patient reflects the delay between a change in the two measurements. A zero total area corresponds to the lack of delay, whereas if heart lags liver, the AUC has a positive sign; and, if liver lags heart, the AUC has a negative sign. The proportion of positive AUCs and their distribution symmetry were compared using Pearson's Chi2 test and the Mann–Whitney U-test. Potential predictors of AUCs positivity were investigated with a binary logistic regression model. Goodness of fit of the model was assessed through Hosmer & Lemeshow test whereas Nagelkerke R2 was used to measure how useful explanatory variables were in predicting the outcome. All statistical procedures were performed using SPSS version 17.0 software package (SPSS Inc., Chicago, Illinois, USA).
have a significant effect on AUC positivity with an O.R. = 20.0 for DFP versus DFO treatment (p = 0.027, Nagelkerke R2 = 0.324). Knowledge of the chelator used is, therefore, essential to describe and compare the speed of decreases in iron loads between heart and liver, as much of this process likely reflects chelating treatment specificities. Finally, to understand if these different temporalities between treatments involved mainly loading or unloading of iron in the small sample under study, we plotted median cardiac R2 in function of median LIC for each treatment group for the first three MRIs (available for all patients in each group and representing 86% of all MRIs). Results are presented in Fig. 2 and show cardiac unloading prevalence in the DFP and combined therapy, while DFO demonstrates a prevalent cardiac iron loading process, whereas all treatments exhibit a prevalence of liver loading process. However the only significant variations were observed between second and third MRI for DFO + DFP on cardiac R2 (p = 0.050; Wilcoxon signed-rank test) and between first and second MRI for DFP on LIC (p = 0.018; Wilcoxon signed-rank test). These data complement previous results by showing that in DFP group liver iron lags cardiac iron in a prevalent unloading process of the heart, showing a pattern of cardiac specificity for DFP, as opposed to DFO. Furthermore, combined therapy demonstrates a pattern of cardiac unloading similar to DFP and a pattern of liver loading similar to DFO.
Results
Discussion
Among the AUCs from the 24 time courses of the patients considered in the study, 9 (37.5%) had an overall negative area (change in liver iron lagging a change in cardiac iron) and 15 (62.5%) had an overall positive area (change in cardiac iron lagging a change in liver iron) (p = 0.221, Pearson's Chi2 test). In this sample, distribution was positively skewed (skewness = 3.3 [Std. Err. = 0.47]; mean = 85.9; median = 25.3) and two large areas were observed in positive direction while AUCs were otherwise similar in both directions. The magnitude of the positive areas averaged 181.7 (median = 87.2; minimum = 12.2; maximum = 1202.5), and the magnitude of the negative areas averaged −73.8 (median = −90.4; minimum = −141.5; maximum = −3.3) resulting in a significant difference between positive and negative ranked AUCs (p b 0.001 Mann–Whitney U-test). Therefore, as in the previously reported work, overall AUCs data (i.e. both their sign, which represent the time delay between organs and their magnitude, which represent how much one organ lags the other) showed that a change in cardiac iron content significantly lags liver iron content [4]. To take in consideration the influence of the chelating treatment on the kinetics of iron loading and unloading in heart and liver, we further analyzed AUCs distribution for each of the three different chelators (Fig. 1). From such analysis, we observed that AUCs clearly behave differently upon treatment (all two-by-two comparisons and single group versus null variations were tested, see Table 2 for details). This finding challenged the first observation on overall AUCs data. To further assess the role of treatments, we developed a binary logistic model for AUC positivity including treatment, gender, time of study, age, cardiac R2* and LIC at first MRI. Treatment was the only variable to
In the present study we tested the hypothesis that, in regularly transfused thalassemia patients, cardiac iron loading or unloading lags
Fig. 1. Box-and-whisker plots of AUCs for each treatment group: central band characterizes the median, box boundaries delimit interquartile range (IQR), whiskers represent 1.5 IQR, whereas outliers are symbolized with circles and extremes (>3 IQRs) with a star. p-value is referred to Mann–Whitney U-test (p = 0.013 not considering the two outliers under DFO treatment).
144
F. Danjou et al. / Blood Cells, Molecules and Diseases 51 (2013) 142–145
Table 2 AUCs magnitude and signs according to chelation treatment. Median
Max.
Min.
AUC differences
Proportion of positive and negative AUCs (Exact test)
Mann–Whitney U-test DFO DFO + DFP DFP
−97.96 DFO vs. DFO + DFP: p = 0.102 115.68 −125.36 DFO + DFP vs. DFP: p = 0.083 21.46 −141.54 –
101.94 1202.45 41.67 −31.42
Wilcoxon signed rank test DFO vs. DFP: p = 0.005 – –
liver iron as formulated by Noetzli et al. [4] In our patient population as a whole, we confirmed a delay in cardiac iron changes. However, the organ-specific mechanisms of iron intake and iron elimination suggested by Noetzli et al. are only one of the key factors able to explain this observation. Although the possible role of genetic factors in organ deposition has been studied [7], in chelated patients, iron overload is usually inversely related to the compliance with chelation. To test the influence of the chelating treatments we selected 24 patients with good compliance with chelation treatment who underwent at least 3 MRIs under the same chelation treatment between 2003 and 2010, and analyzed their heart and liver iron concentration over time. We found that temporality of liver and cardiac iron changes significantly differ between DFO and DFP treatments. In patients treated with DFO, three quarters of AUCs were positive, meaning that cardiac iron significantly lagged liver iron changes, as demonstrated by Noetzli et al. [4] Such result could partly explain why many patients treated with DFO have heart iron overload in absence of relevant liver siderosis [8]. The group of patients treated with DFP showed an opposite pattern, with three quarters of negative AUCs, meaning that liver iron significantly lagged cardiac iron. This is most likely due to a rapid and selective cardiac iron removal in the DFP group, as demonstrated in previous randomized studies comparing the effects of DFP and DFO on heart and liver MRI [9]. These results are further supported by the fact that patients under combined therapy demonstrated, as expected, a pattern that combines both DFO and DFP characteristics (both for AUCs analysis and in term of organs unloading events prevalence). In fact, the association between DFO and DFP has been found able to remove both liver and heart iron in
DFO vs. 0: p = 0.028 DFO + DFP vs. 0: p = 0.401 DFP vs. 0: p = 0.063
DFO vs. DFO + DFP: p = 0.294 DFO + DFP vs. DFP: p = 0.315 –
DFO vs. DFP: p = 0.035 – –
DFO vs. expected: p = 0.039 DFO + DFP vs. expected: p = 0.727 DFP vs. expected: p = 0.453
an additive or synergistic way and more rapidly than the monotherapies [9–11]. It is well-known that body iron is made of diverse pools that are differently regulated, and several studies have already shown that iron chelators do not have the same effect on the different pools of iron [12]. The present data are coherent with such findings showing that DFO and DFP target different iron pools, with DFP targeting more selectively intracellular labile iron responsible of cardiac toxicity and combined therapy affecting different pools of iron at the same time [13–15]. In conclusion, our study clearly shows that the temporality of changes between cardiac and liver iron reflects chelator's access to labile cardiac iron, instead of unrevealing a physiological mechanism. To determine the existence of a physiological temporality between cardiac and liver iron changes, it seems therefore essential to, ideally, study a cohort of non-chelated subjects.
Conflict of interest Authors declare no conflict of interest.
Acknowledgments This research is in part supported by grant: LR.7 Agosto 2007, N. 7, Promozione della Ricerca Scientifica e dell'innovazione Tecnologica in Sardegna, from the Regione Autonoma della Sardegna L.R. 11 1990 and by Telethon (Project number GGP08221).
Fig. 2. The small diamonds, squares and triangles represent the median cardiac R2. The median cardiac R2 is shown as a function of the median liver iron content for each treatment group based on the first three MRIs.
F. Danjou et al. / Blood Cells, Molecules and Diseases 51 (2013) 142–145
Appendix A. Supplementary data Supplementary data to this article can be found online at http:// dx.doi.org/10.1016/j.bcmd.2013.06.001.
References [1] B. Modell, M. Khan, M. Darlison, et al., Improved survival of thalassaemia major in the UK and relation to t2* cardiovascular magnetic resonance, J. Cardiovasc. Magn. Reson. 10 (2008) 42. [2] D.J. Pennell, T2* magnetic resonance and myocardial iron in thalassemia, Ann. N. Y. Acad. Sci. 1054 (2005) 373–378. [3] L.J. Anderson, S. Holden, B. Davis, et al., Cardiovascular t2-star (t2*) magnetic resonance for the early diagnosis of myocardial iron overload, Eur. Heart J. 22 (2001) 2171–2179. [4] L.J. Noetzli, S.M. Carson, A.S. Nord, T.D. Coates, J.C. Wood, Longitudinal analysis of heart and liver iron in thalassemia major, Blood 112 (2008) 2973–2978. [5] J.C. Wood, C. Enriquez, N. Ghugre, et al., Mri r2 and r2* mapping accurately estimates hepatic iron concentration in transfusion-dependent thalassemia and sickle cell disease patients, Blood 106 (2005) 1460–1465. [6] M.A. Tanner, T. He, M.A. Westwood, D.N. Firmin, D.J. Pennell, Multi-center validation of the transferability of the magnetic resonance t2* technique for the quantification of tissue iron, Haematologica 91 (2006) 1388–1391.
145
[7] R. Origa, S. Satta, G. Matta, R. Galanello, Glutathione s-transferase gene polymorphism and cardiac iron overload in thalassaemia major, Br. J. Haematol. 142 (2008) 143–145. [8] M.A. Tanner, R. Galanello, C. Dessi, et al., Myocardial iron loading in patients with thalassemia major on deferoxamine chelation, J. Cardiovasc. Magn. Reson. 8 (2006) 543–547. [9] M.A. Tanner, R. Galanello, C. Dessi, et al., A randomized, placebo-controlled, double-blind trial of the effect of combined therapy with deferoxamine and deferiprone on myocardial iron in thalassemia major using cardiovascular magnetic resonance, Circulation 115 (2007) 1876–1884. [10] R. Galanello, A. Agus, S. Campus, et al., Combined iron chelation therapy, Ann. N. Y. Acad. Sci. 1202 (2010) 79–86. [11] P. Evans, R. Kayyali, R.C. Hider, J. Eccleston, J.B. Porter, Mechanisms for the shuttling of plasma non-transferrin-bound iron (NTBI) onto deferoxamine by deferiprone, Transl. Res. 156 (2010) 55–67. [12] J.B. Porter, Optimizing iron chelation strategies in beta-thalassaemia major, Blood Rev. 23 (Suppl. 1) (2009) S3–S7. [13] L.J. Anderson, B. Wonke, E. Prescott, et al., Comparison of effects of oral deferiprone and subcutaneous desferrioxamine on myocardial iron concentrations and ventricular function in beta-thalassaemia, Lancet 360 (2002) 516–520. [14] C. Hershko, G. Link, A.M. Konijn, Z.I. Cabantchik, Objectives and mechanism of iron chelation therapy, Ann. N. Y. Acad. Sci. 1054 (2005) 124–135. [15] D.J. Pennell, V. Berdoukas, M. Karagiorga, et al., Randomized controlled trial of deferiprone or deferoxamine in beta-thalassemia major patients with asymptomatic myocardial siderosis, Blood 107 (2006) 3738–3744.