Page 135 - Haematologica March 2020

Iron chelation in lower risk MDS patients
The mechanisms by which ICT influences OS after a relatively short exposure to iron chelation therapy (medi- an duration of 13 months) are not completely understood. A recently published study of the EUMDS Registry, as well as the follow-up data of this study, demonstrated detectable labile plasma iron (LPI) levels to be associated with inferior OS in LR-MDS patients.31,32 The risk of dying prematurely in patients with detectable LPI levels occurred too early in this study to explain this risk by classical IOL due to organ toxicity (e.g. liver and heart) after long-term transfusions, but this indicates a direct toxic effect associated with elevated LPI levels.31
Likewise, there is increasing evidence that increased LPI levels may be a general predictor of an increased non- relapse mortality during and after hematopoietic stem cell transplantion.33
Toxic iron species are known to catalyze the cellular generation of ROS, which play a key role in cellular dam- age.34,35 ROS damage (mitochondrial) DNA, with poten- tial consequent genomic instability, mutagenesis, and cell death. ROS are associated with leukemic transformation of the MDS clone.6 Moreover, ICT is associated with a decrease in LPI and ROS.6,16 Overall, the present study indicates that ICT may partly counteract the unfavorable consequences of secondary IOL.
In up to 31.0% of chelated patients a reduction in trans- fusion density was observed during at least one interval between check-ups. Likewise, 27.4% of the responding patients became, at least temporarily, transfusion inde- pendent. Platelet responses were less frequently observed. However, platelet count in this context was less relevant because the platelet counts in both groups were within the normal range, and will not likely lead to severe bleeding complications. Contemporary treatment with ESA and/or lenalidomide may have enhanced these responses.
Several previous studies recorded hematologic respons- es to ICT.11-16 While the percentage of patients with hema- tologic responses in these studies are in line with the present study, none of the former studies included a con- trol group in their analyses. One of the key factors is the relatively short period of ICT (median 13 months) in this study. The duration of ICT may improve by the introduc- tion of a better tolerated formulation of deferasirox.36 ICT is usually prescribed relatively late after detection of signs of IOL. Earlier initiation of ICT may prevent or decrease the occurrence of transfusional iron toxicity on hematopoiesis. Moreover, we recorded data only at 6- monthly intervals. Short duration hematologic responses in between check-ups may be missed by this approach. But on the other hand, short-term responses may not be clinically relevant.
Pre-clinical studies have shown a beneficial effect of ICT on hematopoiesis.35,37 Inhibition of the transcription factor NF-κB, involved in many cellular processes, and modulation of mammalian target of rapamycin (M-TOR) signaling, a major regulator of cell death and proliferation, have been proposed to play a role.17 Future studies should address this issue appropriately.
In the trajectory analyses, ferritin responses occurred in up to 23.5% of the chelated patients. Serum ferritin levels have frequently been reported to be a prognostic marker in LR-MDS patients, but serum ferritin is an imprecise surrogate marker for secondary IOL and toxicity.38,39 This is reflected by the observation that a relatively small pro-
portion of chelated patients have a considerable decrease in serum ferritin levels, while these patients show a sig- nificant survival benefit. Serum ferritin levels are influ- enced by the stage of MDS and by concurrent infection and inflammation, which is common in LR-MDS patients.38 In addition, there is no convincing evidence regarding its use for monitoring secondary IOL in MDS patients.38,40 Tissue biopsy and MRI T2* are currently regarded to be the most specific and sensitive diagnostic tests for detecting IOL.38 However, the clinical utility of these assays remains unclear in MDS and invasiveness (biopsy), unavailability, and expense (MRI T2*) hamper their general use in clinical practice. As discussed above, LPI is associated with inferior survival in LR-MDS patients.31,32 Future studies are warranted to evaluate the effect of ICT on LPI levels as a measure of iron toxicity. Measurement of oxidative stress, including malondialde- hyde, a long-lasting lipid peroxidation product, formed as a consequence of oxidative stress from IOL, are also pos- sible future markers to detect and monitor the biological consequences of secondary IO in LR-MDS patients, should they be proven to correlate with clinical out- comes.41,42
Analysis of renal function demonstrated that ICT is associated with an increase in creatinine levels. In some patients, this will be clinically relevant and/or a reason to stop or lower the dose of ICT. In other patients, an increase in creatinine levels will not affect cessation of ICT.
This large cohort, with prospectively collected ‘real-life’ data from diagnosis, provides a unique opportunity to study the effect of ICT in a large number of lower-risk MDS patients in daily practice. An important strength of the study is that the results can be widely generalized to this, mostly elderly, patient population with multiple comorbidities, who are typically excluded from clinical trials. The variation in iron chelation practice across the different countries, due to variable interpretation of the poor quality outcome data for ICT in MDS, made it pos- sible to compare the effects of ICT on OS to a non-chelat- ed control group. In Europe, unlike in the United States, socio-economic status does not influence the prescription of ICT (either deferoxamine or deferasirox) because the costs are covered by the national health systems.
Since conventional statistical modeling is limited by the number of co-variates to be added to a model, propensi- ty-score matched analysis is able to incorporate more confounding factors in the model, including country-spe- cific effects. Confounding by indication, a common prob- lem in observational studies, is maximally reduced by using the propensity-score matched method and is, there- fore, a major strength of this study. To our knowledge, we are the first to apply this method in order to adequate- ly deal with confounding in this setting.
Limitations of our study include the moderate sample size of the deferoxamine and deferiprone groups. In addi- tion, our analysis could not consider differences in dosing schemes and therapy compliance. This prevented us from drawing definitive conclusions of the effect of the sepa- rate iron chelators on OS. Data were collected at the scheduled 6-monthly intervals. Subtle changes in patient- related factors in the intervening 6-month period may have been missed. Not all patients can be matched by the propensity score approach. This might introduce selec- tion bias. However, the same magnitude and direction of
haematologica | 2020; 105(3)
649