Oral chelators deferasirox and deferiprone for transfusional iron overload in thalassemia major: new data, new questions

In this issue of Blood, there are reports of 2 randomized clinical trials and a large retrospective epidemiology study that represent crucial advances in the field of oral agents to treat transfusional iron overload in thalassemia.^1-3  Here I will summarize the results of these studies, point out some of their strengths and weaknesses, and propose one hematologist's view of where this work leaves the field.

Deferoxamine (DFO; Desferal and generic) has been the standard iron chelator since the 1970s. DFO is both safe and effective for transfusional hemosiderosis. A hexadentate chelator, it binds iron tightly, and the iron-DFO complex is excreted in both urine and stool. DFO is administered as long parenteral infusions because the plasma half-life is short (minutes) and it is not active orally. Thus, it is given as an overnight subcutaneous infusion 5 to 7 nights/wk. The DFO-iron chelate is charged and does not readily enter and leave cells.⁴  Parenteral administration and the daily nuisance of an infusion pump hinder optimal compliance. Nevertheless, in the DFO era, over the past generation, dramatic strides in survival of thalassemia patients have occurred.⁵ 

Deferiprone (Ferriprox and others) is an orally active hydroxypyridineone first used in humans in 1987. Deferiprone is a bidentate chelator (3 molecules surround one iron ion). An advantage of this compound is that the iron(III) chelate of deferiprone carries no net charge and therefore can penetrate membranes easily, allowing removal of potentially toxic iron from tissues.⁶  Many Blood readers are aware of a controversy over the safety of deferiprone that arose in the late 1990s because of an observation of hepatic fibrosis during a clinical trial.⁷  However, in subsequent studies this problem has not been a significant toxicity issue for deferiprone.⁸  The history of deferiprone and this safety debate were well summarized in a 2003 Blood “Perspective.”⁹  Deferiprone often causes gastrointestinal symptoms. Idiosyncratic side effects that are potentially severe include erosive arthritis (common in patients in South Asian countries, from 5% to > 20%) and neutropenia (up to 5% of patients), including severe agranulocytosis (up to 0.5% of patients); close monitoring is required. Typical dosage for deferiprone is 75 mg/kg/d in 3 divided doses, up to 100 mg/kg daily.^3,10 

Deferasirox (ICL670, Exjade) belongs to a new class of oral tridentate chelator, N-substituted bis-hydroxyphenyltriazoles. Deferasirox, the result of a concerted discovery program, underwent extensive safety testing and clinical trials including preclinical studies,¹¹  initial phase 1 and iron balance studies,¹²  phase 2 efficacy studies in adult¹³  and pediatric¹⁴  thalassemia patients, patients with a variety of anemias or unable/noncompliant with DFO,¹⁵  and the phase 3 clinical trial discussed here.¹  With a plasma half-life of 8 to 16 hours, once-daily dosing permits circulating drug at all times to scavenge non–transferrin-bound “labile plasma iron,”¹⁶  the chemical species responsible for tissue damage in iron-overloaded subjects, by means of toxic oxygen intermediaries.¹⁷  Deferasirox-iron complexes are excreted in the stool.

The 3 compounds are compared in Table 1 to properties of an ideal chelator. One “theoretical” advantage with real clinical import is access of chelators to intracellular iron, particularly in cardiac myocytes. In cultured heart muscle, deferasirox and deferiprone have rapid access to intracellular iron pools, whereas DFO does not.¹⁸  Nevertheless, high-dose continuous DFO, administered via central catheter, can dramatically reverse cardiac toxicity of iron overload.⁴  High-dose continuous DFO has been the “standard” initial therapy for cardiac iron overload.

Trials to test chelator effectiveness necessarily require a choice of a primary outcome measure. Potential “hard” end points or surrogates are summarized in Table 2. Different outcomes were chosen for the 3 studies discussed here. (1) Hepatic iron content (HIC) by liver biopsy was used for the deferasirox-versus-DFO trial.¹  Liver biopsy has been the “gold standard” for iron balance studies, but the technique is invasive, expensive, and subject to variability within and between research subjects. (2) The epidemiologic study of deferiprone examines onset of symptomatic cardiac disease as the primary outcome measure.²  This is a compelling end point because iron-related cardiac disease is a major clinical problem. (3) For their comparative trial of DFO and deferiprone, Pennell and colleagues³  used the cardiac magnetic resonance imaging (MRI) parameter T2* (pronounced “T-2-star”). Iron concentration in tissues (in this case, the myocardium of the cardiac septum) is inversely related to T2*. Low T2* values (< 8 ms compared with normal [> 20 ms]) are related to risk of heart failure and death in iron-overloaded thalassemia patients.²⁰  T2* reports initially raised concerns about whether it is a good surrogate for cardiac iron concentration²⁵  for technical reasons as well as the paradox that liver iron correlated poorly with cardiac T2*. These points have largely been settled. In an iron-overload gerbil model, T2* is indeed inversely related to myocardial iron measured directly.²⁶  The poor correlation of T2* with hepatic iron is now understood on a kinetic basis. The liver can be readily unloaded by aggressive chelation much more rapidly than can the heart, giving rise to patients with low T2* values (heart still iron loaded) while the liver has been unloaded to values previously thought of as “safe.”²⁷ 

Pennell et al³  compared oral deferiprone to subcutaneous DFO in 61 thalassemia major patients without symptomatic heart failure. The entry criteria required T2* to be abnormal (< 20 ms) but not “severe” (< 8 ms) and left ventricular ejection fraction (LVEF) to be greater than 56%. Technical details of the single-breath-hold T2* method were previously published.²⁸  Superconducting quantum interference device (SQUID; Table 2) was used to measure HIC at baseline and 12 months. Compliance was closely monitored. The primary outcome measure was change in myocardial T2* at 6 and 12 months. The rate of rise in T2* was significantly higher to deferiprone (mean dose 92 mg/kg/d) than to DFO (mean dose of 43 mg/kg, 5.7 d/wk). In addition, there was a greater increase in LVEF in the deferiprone group compared with the deferoxamine group, all within the “normal” range. Mean HIC fell slightly in each group. Gastrointestinal symptoms and joint pain were common side effects. This prospective trial validates the prior retrospective series of Anderson et al²⁹  that had been controversial.²⁵  A potential weakness of the trial is the use of the surrogate T2* as primary outcome, because the clinical significance of a few more milliseconds within the mildly abnormal range is unknown if taken alone. However, taken together with in vitro data and with the epidemiologic data of Borgna-Pignatti et al,²  this trial provides important further evidence that, despite the concerns about deferiprone chelation efficiency based on 3:1 stoichiometry, deferiprone is probably better in vivo at improving myocardial iron than DFO (at the doses used).

Phase 3 trial results for deferasirox in a randomized controlled comparison trial with DFO are reported by Cappellini and colleagues.¹  The trial involved nearly 600 patients (half were 16 years of age or younger) with transfusion-dependent β-thalassemia. The study was designed as a noninferiority trial, measuring hepatic iron by biopsy at baseline and after one year of therapy. Success was defined as either maintenance or improvement in HIC (depending on the baseline levels). Based on the study results, the United States Food and Drug Administration (FDA) approved the drug for transfusional iron overload for patients older than 2 years of age in November 2005. The drug is still under regulatory review in Europe.

This report deserves careful scrutiny from hematologists who care for patients with thalassemia and other disorders of transfusional iron overload. The dose choices for DFO and for deferasirox were based solely on baseline HIC at study entry. The decision to use this baseline value and the choice of what are now known to be relatively low deferasirox doses caused the study to fail to meet its overall primary end point (ie, at low doses of deferasirox, 5-10 mg/kg/d, increased HIC was observed). However, at doses of 20 to 30 mg/kg/d, the doses for HIC greater than 7 mg/g dry weight, noninferiority of deferasirox compared with DFO was established, with 60% versus 59% achieving a successful outcome, respectively. Assessed by ferritin concentration, 20 mg/kg/d deferasirox was sufficient to maintain mean ferritin levels over 52 weeks, whereas 30 mg/kg yielded a reduction in ferritin level. In retrospect, several design decisions contributed to the mixed outcome of “failure” at low doses and success at higher doses. For example, results from an earlier phase 2 trial of deferasirox in children,¹⁴  which also used low doses that proved ineffective, were not available in time to guide the phase 3 trial. As well, additional factors proved important in deferasirox efficacy. For example, transfused iron burden was shown to be a strong predictor of the iron response at a year.^30,31  Likelihood of failing to achieve iron balance was much higher for patients receiving more than 0.5 mg/kg/d iron.³¹  Adverse drug reactions in deferasirox trials have included modest rise in creatinine level, rarely clinically significant. Increased transaminases were observed occasionally. Common side effects of deferasirox include transient gastrointestinal symptoms in 15% and rash in 11%.

Borgna-Pignatti and colleagues²  present an interesting retrospective cohort study of cardiac events in Italian thalassemia patients who switched from DFO to deferiprone. The study included more than 500 patients with thalassemia major at 7 large Italian centers who were alive in 1995 and who had neither prior heart disease nor subsequent marrow transplantation. Deferiprone was given to 157 subjects during some period in the subsequent 9 years, whereas 359 received only DFO. Many of the deferiprone patients were initially part of clinical trials, before the drug became available on a compassionate-use basis in 1997 and commercially in 2000. In a time-to-event analysis, the authors noted 52 cardiac events, including 10 deaths, in patients while on DFO and 0 events while on deferiprone. The result is dramatic and unexpected, whether one believes that deferiprone would have been given mostly to DFO-intolerant patients or not. Although potential bias could easily arise in a retrospective study of unmatched groups, the authors have examined possible biases in a comprehensive fashion, controlling for as many as possible, and explaining the rest with admirable clarity and near-perfect patient ascertainment. This stunning finding, coupled with similar but less rigorous data from other sites, is hard to ignore. The results confirmed a smaller retrospective analysis of Piga et al.³² 

Perfect and unbiased cost comparisons among commercial versions of all 3 drugs are not yet possible for any single country, and national price differences abound. Although a rough first-order approximation of price ranking might be deferiprone < deferoxamine < deferasirox, this may not be true in every country (Table 3 provides some comparisons of available costs). Nondrug obligatory costs are important as well: deferiprone therapy requires weekly complete blood count (CBC)/differential count; DFO requires ancillary supplies for infusion. Deferasirox will cost more than twice as much as DFO in the United States. At least until deferasirox, chelators heretofore have not been the major cost of caring for American thalassemia patients; this dubious honor goes to the transfusions themselves.³⁵  But the cost of deferasirox will be a significant new burden even in developed nations with strong health insurance programs, and it will be prohibitive in the developing world, without substantial discounts. An initial attempt at cost-effectiveness analysis was presented in abstract form at the American Society of Hematology 47th Annual Meeting in 2005, wherein the extra cost of the medication was weighed (favorably) against the cost of illness and death from noncompliance and iron overload for DFO.³⁶  This analysis was sponsored by the manufacturer.

More than 20 deferasirox abstracts were presented at the 2005 American Society of Hematology Annual Meeting. These included clear evidence from both phase 2 and phase 3 clinical trials that transfusional iron loading (expressed as mg/kg/d of transfused iron) had a dramatic effect on the ability of deferasirox doses to maintain or reduce hepatic iron^30,31  and presentation of the randomized trial of deferasirox versus DFO in sickle cell disease.³⁷  Porter et al³⁸  demonstrated improved cardiac T2* with deferasirox in patients from their site in the phase 2 and 3 trials. Molar efficiency of DFO and deferasirox was compared by examining net iron balance as a function of input transfusional iron and chelator dose.³⁹  These results have important implications to consider along with the phase 3 trial¹ : chelator efficiency and clinical effectiveness in vivo are not a function of chelate stoichiometry alone. Iron removal also depends on achievable plasma concentration, host factors, degree of loading, and rate of accessibility of stored iron to chelator.

For treaters and patients, a prudent path would be first to take to heart the main messages of each of the 3 studies. I believe these are the take-home messages. (1) Cappellini et al¹  show that at 20 to 30 mg/kg/d, deferasirox can keep most but not all patients in even or negative iron balance in rough equivalence with moderate doses of DFO. (2) Pennell et al³  prove that deferiprone was able to improve not only T2* in asymptomatic patients, but these investigators also provided circumstantial evidence that their patients with “normal” LVEF had subclinical disease, based on improvements in LVEF with chelators. Pennell and colleagues have also initiated a randomized, prospective trial of DFO plus placebo versus DFO plus deferiprone in borderline T2* status; results should soon be published. (3) Borgna-Pignatti et al²  provide the strongest evidence to date that deferiprone should be considered cardioprotective in comparison with DFO, albeit in a retrospective setting without formal matching. But these 3 studies have not yet addressed some crucial questions.

What is the optimum deferasirox dose for patients with high iron intake? Some patients with high transfusion burdens will probably have rising ferritin levels and HICs if treated with the approved doses of 20 to 30 mg/kg/d.

If so, alternative strategies will be obligatory. These might include higher daily doses (which may cause more diarrhea¹² ), twice-daily dosing, or possible combination regimens with DFO. The latter option would offer lower average price than deferasirox alone (and ideally better net iron chelation) but would lose the advantages of an oral drug. A combination of the 2 oral drugs, deferasirox and deferiprone, requires detailed study before it can be safely recommended.

What is the proper role for deferiprone? Based on side-effect profile alone, in my opinion, deferiprone is likely to remain a “second-line” drug to deferasirox and DFO.

This is not to say that deferiprone will not have a crucial role. In particular, deferiprone plus DFO in some combination appears to be the emerging “treatment of choice” for significant cardiac dysfunction from iron overload, as in vivo⁴⁰  and in vitro⁴¹  evidence support the biochemical rationale.⁴²  Deferiprone is currently available only on a compassionate-use basis in the United States. If deferasirox is successfully used where patients used to be noncompliant with DFO, its main success will be to prevent the need for such rescue therapy altogether.

What shall we tell our patients? In our thalassemia center, we are presenting the approval of deferasirox as a major advance and treatment option.

After politely receiving this information, however, many DFO users with good compliance and HICs in safe ranges are choosing not to switch until more data become available, because they are doing fine with their pumps and long-term deferasirox data are not yet available. We have not recommended deferasirox to patients with overt myocardial dysfunction because no published data support this use. We now adamantly suggest deferasirox to DFO-noncompliant patients without symptomatic heart disease.

Our center is participating in a National Institutes of Health (NIH)–sponsored randomized trial comparing intensive DFO plus deferiprone to DFO plus placebo for patients with low LVEF or symptomatic heart disease, and we tell such patients that continuous-infusion DFO is standard care but that addition of deferiprone on study (or for nonstudy participants as compassionate use) may be advantageous. Additional carefully designed studies are required to answer pressing questions about many drug combinations and treatment scenarios. Examples include both small-scale trials (eg, pilot pharmacodynamic studies of the 3 drugs in various combinations) and large-scale efforts (eg, initiation of a randomized, prospective phase 3 trial comparing deferasirox and deferaprone to assess relative safety, efficacy, and cardioprotection).

We look forward to a future where complications due to iron overload are rare, and our patients certainly look forward to life without “the pump.”