Why measure serum ferritin?

Serum ferritin (SF) generally correlates with body iron stores, and is relatively easy and inexpensive to determine repeatedly. Serum ferritin is most useful in identifying trends. A decreasing trend in SF is good evidence of decreasing body iron burden but absence of a decreasing trend does not exclude a decreasing iron burden. However, an increasing SF trend implies an increasing iron burden but may also be due to inflammation or tissue damage, so clinical judgment must be used to interpret these trends. Long term control of SF is also a useful guide to the risk of complications from iron overload in TM; many studies have shown an association between the control of serum ferritin and prognosis (Borgna-Pignatti 2004, Davis 2004, Gabutti 1996, Olivieri 1994). Studies have identified a significantly lower risk of cardiac disease and death in at least two-thirds of cases where serum ferritin levels have been maintained below 2,500 µg/L (with Deferoxamine, or DFO) over a period of a decade or more (Olivieri 1994). Observations with larger patient numbers show that maintenance of an even lower serum ferritin of 1,000 µg/L may be associated with additional clinical advantages (Borgna-Pignatti 2004) (see Table 2).

Table 2.

Use of SF for monitoring chelation treatment.

What are the limitations of serum ferritin measurements?

Most SF assays were developed mainly for detecting iron deficiency, and the linear range of the assay at high SF values needs to be known. SF must be performed in a laboratory that has established how to dilute samples with high values, to give readings within the linear range of the assay. SF measures do not always predict body iron or trends in body iron accurately. In TM, variation in body iron stores accounts for only 57% of the variability in plasma ferritin (Brittenham 1993). This variability is in part because inflammation increases serum ferritin, and partly because the distribution of liver iron between macrophages (Kupffer cells) and hepatocytes in the liver has a major impact on plasma ferritin. A sudden increase in serum ferritin should prompt a search for hepatitis, other infections, or inflammatory conditions.

A lack of fall in SF with chelation does not therefore necessarily prove that the patient is a ‘non responder’ to the chelation regime. As outlined above, this can be because inflammation may have falsely raised SF, or because the relationship between body iron and SF is not always linear, particularly in the context of inflammation or tissue damage (Adamkiewicz 2009), and body iron can fall considerably from a high starting point (e.g. LIC >30 mg/g dry wt) before a change in ferritin is clear. Below 3000 µg/L SF values are influenced mainly by iron stores in the macrophage system, whereas above 3000 µg/L they are determined increasingly by ferritin leakage from hepatocytes (Davis 2004, Worwood 1980). Day-to-day variations are particularly marked at these levels. The relationship between serum ferritin and body iron stores may also vary depending on the chelator used (Ang 2010) and by duration of chelation therapy (Fischer 2003).

Uses for liver iron concentration monitoring

• - To identify whether body iron is adequately controlled.
  Adequate control of LIC is linked to the risk of hepatic damage as well as the risk of extrahepatic damage. Normal LIC values are up to 1.8 mg/g dry wt, with levels of up to 7 mg/g dry wt seen in some non-thalassaemic populations without apparent adverse effects. Sustained high LIC (above 15-20 mg/g dry wt) have been lined to worsening prognosis, liver fibrosis progression (Angelucci 1997) or liver function abnormalities (Jensen 2003). In the absence of prior iron chelation therapy, the risk of myocardial iron loading increases with the number of blood units transfused and hence with iron overload (Jensen 2003. Buja 1971). However, the relationship between LIC and extra-hepatic iron is complicated by chelation therapy as iron tends to be accumulate initially in the liver and later in the heart but also is removed more rapidly from the liver than the heart by chelation therapy (Noetzli 2008, Anderson 2004). Thus in patients receiving chelation therapy, whilst high LIC increases the risk of cardiac iron overload, the measurement of LIC will not predict myocardial iron and hence cardiac risk reliably, and myocardial iron may be found in some patients despite currently well controlled LIC.
• - To determine iron balance: is body iron increasing or decreasing on current therapy?
  LIC is the most reliable indicator of body iron load, which can be derived from the following formula: Total body iron stores in mg iron /kg body wt = 10.6 × the LIC (in mg/g dry wt) (Angelucci 2000). Sequential measurement of LIC is the best way to determine whether body iron is increasing or decreasing with time (iron balance). While serum ferritin is simple, relatively inexpensive and can be repeated frequently, LIC determination should be considered for those patients whose serum ferritin levels deviate from expected trends (i.e. those with suspected co-existing hepatitis, or patients on chelation regimens with variable or uncertain responses), as this may reduce the risk of giving either inadequate or excessive doses of chelation therapy. Since the relationship of SF to iron overload and iron balance has not yet been established, assessment of LIC may be particularly useful when new chelating regimes are being used. At high levels of SF (>4000 µg/L), the relationship to LIC is not linear and patients may show fall in LIC (negative iron balance) without a clear trend in SF in the first 6-12 months. When a patients fails to show a fall in SF over several months the change in LIC can identify whether the current regime is adequate or need to be modified (increased frequency or adherence, increased dose, or change in regime).

Methods for measuring LIC

• - Biopsy
  Measurement of LIC was initially done by chemical determination on a liver biopsy sample (fresh, fixed or from dewaxing of paraffin-embedded material) (see Table 3). Biopsy is an invasive procedure, but in experienced hands has a low complication rate (Angelucci 1997). Inadequate sample size (4 mg dry wt or about a 2.5 cm core length) or uneven distribution of iron, particularly in the presence of cirrhosis (Villeneuve 1996), may give misleading results however. Biopsy also allows the evaluation of liver histology which cannot yet be reliably estimated by non-invasive means. Laboratory standardization is not trivial and differences between laboratories, for example in wet to dry weight ratios, can mean that results from different labs may not be equivalent.
• - SQUID
  Magnetic biosusceptometry (SQUID) (superconducting quantum interference device) determines the paramagnetism of the liver which is proportional to LIC (Brittenham 1993). Current methodology requires liquid helium which is very expensive. Furthermore, the SQUID apparatus needs to be in an environment away from paramagnetic forces (e.g. lifts, cars) which is often impractical. For these reasons, the current generation of SQUID devices are unlikely to be used outside a small number of well-resourced centres. Surprisingly not all SQUID devices have been calibrated the same, so comparison of results from different centres must be interpreted with caution unless the relevant machines have been cross-validated.
• - MRI
  MRI techniques are now becoming the most widely used methods for LIC determination. The first techniques compared the signal in the liver or heart with that of skeletal muscle, which does not accumulate iron (Jensen 1994). However, this is not widespread use today and has been superseded by better methods. The principle shared by all MRI techniques currently used is that when a radio-frequency (rf) magnetic field pulse is applied to the tissue (e.g. liver or myocardium), protons take up energy, altering their spin orientation and they later relax returning to their original state. With spin echo, after the pulse the nuclei take time to relax in the “relaxation time”; T1 in the longitudinal plane, and T2 in the transverse plane. Values may also be expressed as relaxation rates, the R1 rate (is the same as 1/T1) and the R2 rate (is the same as 1/T2). A variation of this principle are Gradient Echo techniques, achieved by applying a strong graded magnetic field to the radio-frequency (rf) pulse that is used for spin echo. This relies on multiple echoes over a shorter acquisition time period than spin echo techniques. The shorter acquisition may improve sensitivity and can be measured as T2* (in ms), where 1/T2* = 1/T2 + 1/T2’, and T2 is the tissue relaxation time and T2’ is the magnetic inhomogeneity of the tissue. An important point is that tissue iron concentration is not linearly related to T2* or the T2, but is linearly related to 1/T2* or 1/T2 (R2* or R2). Both gradient and/or spin echo techniques have been used in clinical practice. T2* (or R2*) can be achieved with a single breath hold, while T2 or R2 take a little longer to acquire data. Manufacturers of suitable MRI scanners are: Siemens (Erlangen, Germany); GE Healthcare (Milwaukee, WI, USA); Philips Healthcare. The strength of the magnetic field applied by these scanners is measured in Tesla (T) units. Most imaging is done on 1.5T machines but 3T machines give a better signal to noise ratio. However, 3T machines have greater susceptibility to artefacts, and the maximum detectable iron level is also halved (which is too low for many patients) (Wood 2008, Storey 2007). At present only 1.5T machines are widely used with reliable precision and accuracy based on standardised validation procedures. Liver packages (including standard sequences and analysis of the data) are included in the software provided with these MRI machines. Specialized LIC analysis software can also be bought separately.
  A note of caution is that the different MRI techniques may not be equivalent – at least in the manner they are currently calibrated and practiced. The first widely used technique was the T2* technique (Anderson 2001), where liver biopsy was used to calibrate the method. Although this demonstrated the principle of T2* to measure liver iron, unfortunately due to factors such as long echo times (TE 2.2-20.1 ms), and multi breath-hold acquisition, the calibration differs from later techniques, and can underestimate LIC by two-fold. Therefore studies using this calibration may underestimate LIC (Garbowski 2009). The R2 based Ferriscan technique appears to have acceptable linearity and reproducibility up to LIC values of about 30 mg/g dry wt (St Pierre 2005), with an average sensitivity of >85% and specificity of >92% up to an LIC of 15 mg/g dry wt, and has been registered in the EU and US. For calibration of Ferriscan, the MRI machine must use a Phantom supplied by the company, while the data acquired is sent via internet for analysis by dedicated Ferriscan software (payment per scan analyzed). A particular advantage of this technique is that it can be applied with little training, at any centre with a reasonably up-to-date MRI machine (see Table 3).

Table 3.

Rationale, advantages and disadvantages of LIC determination by MRI and biopsy.

Evidence of beneficial effects

DFO was the first chelator introduced clinically. A large body of literature has since emerged on the changing complications and improved survival, predominantly from retrospective cohort analysis. As no treatment alternatives were available at the time of its introduction, the benefits of its long term use are clearer than for newer chelators, where patients have often received more than one chelation treatment over a lifetime. The main disadvantages of the treatment are that it is costly and it must be administered parenterally which is uncomfortable and time consuming. Also because of its short half live it typically only chelates iron during the time infused, therefore leaving 12 or more hours with no active chelator with standard regimens. The increased toxicity of DFO at low levels of body iron (see Appendix 2) means that guidelines for its use have been conservative, generally recommending that therapy not be started until SF levels reach 1000 µg/L, and with care to avoid over-chelation below this SF value.

Effects on serum ferritin

Long term control of SF has been linked to protection from heart disease and to improved survival if levels are consistently less than 2500µg/L (Olivieri 1994) with even better outcomes at levels <1000µg/L (Borgna-Pignatti 2004). Four decades of clinical experience clearly show that ferritin can be controlled with DFO monotherapy at 40-50 mg/kg administered as an 8-10-h infusion at least 5 times a week. In children however, mean daily doses should not exceed 40 mg/g because of the effects on growth and skeletal development. Guidelines about the dosing required to control iron overload were based on retrospective data until recently. A randomised study in 290 TM patients identified the doses required to stabilize or decrease SF, with a mean daily dose of 42 mg/kg resulting in a small decrease in serum ferritin of 364µg/L at one year, whereas 51 mg/kg resulted in an average decrease of approximately 1,000 µg/L (Cappellini 2006). Further analysis shows that response is also linked to the transfusion rate and that larger doses are required in patients requiring higher transfusions (see below) (Cohen 2008). Thus the effectiveness of DFO at controlling SF is related to dose, frequency and duration of exposure and transfusion rate.

Effects on liver iron

Administered at least 5 times a week in sufficient doses, DFO is effective in controlling liver iron and hence total body iron stores (Brittenham 1993). In a prospective randomized study (Cappellini 2006), a mean dose of 37 mg/kg stabilised LIC for patients with baseline LIC values of between 3 and 7 mg/g dry wt. For patients with LIC values between 7 and 14 mg/g dry wt, a mean dose of 42 mg/kg resulted in a small decrease of 1.9 mg/kg dry wt over a 1 year interval. In patients with LIC values >14 mg/g dry wt, a mean dose of 51 mg/kg resulted in LIC decreases of an average of 6.4 mg/g dry wt. Thus a dose of 50 mg/kg at least 5 days a week (giving a mean daily dose of 50 × 5/7 = 36 mg/kg) is recommended if a significant decrease in optimal LIC levels is required (see above). It should be emphasised that these are average changes and that the dose required may increase or decrease depending on transfusion requirements (Cohen 2008).

Effects on heart function

Subcutaneous therapy has long been known to prevent (Wolfe 1985) or improve asymptomatic cardiac disease in thalassaemia major (Aldouri 1990, Freeman 1983). After the introduction of DFO, the incidence of iron- induced heart disease in different cohorts of patients fell progressively – with a key factor being the age of starting treatment (Borgna-Pignatti 2004, Brittenham 1994). Symptomatic heart disease can be reversed by high dose intravenous treatment (Davis 2000, Cohen 1989, Marcus 1984). The same results can be obtained with excellent long-term prognosis with lower doses (50-60 mg/kg/day – see below), and consequently less drug toxicity using continuous dosing (Davis 2004, Davis 2000). Continuous intravenous doses of 50-60 mg/kg/day typically normalise LVEF in a period of three months (Anderson 2004), significantly before liver or heart iron stores have been normalised. However, if advanced heart failure has developed before treatment is intensified, the chances of successful rescue are reduced. Early intervention for decreased LV function is therefore recommended. Once heart function has improved, sustained compliance is critical to improve outcomes, especially while myocardial iron remains increased (Davis 2004).

Effects on heart iron (mT2*)

Myocardial iron can improve with either subcutaneous or intravenous therapy provided treatment is given in adequate doses and frequency. Improvement in mild to moderate cardiac T2*, even at low intermittent doses (5 days a week) has been confirmed by prospective randomised studies (Pennell 2014, Pennell 2006b, Tanner 2006). For patients with established mild to moderate myocardial iron, a simple increase in dose or frequency of use may be sufficient to improve the mT2*. For example at relatively low doses of 35 mg/kg, an average improvement in T2* of 1.8 ms over one year has been shown (Pennell 2006b). At a slightly higher dose of 40-50 mg/kg five days a week, patients showed an improvement of 3 ms over one year (Porter 2005a). When mT2* is < 10 ms, as with other iron chelators, it will take several years of sustained and compliant therapy to normalise myocardial iron (Porter 2002). For T2* values <10 ms, a simple 5 day a week s.c. DFO at standard doses is unlikely to be sufficient, and treatment intensification is indicated. This could involve higher dose continuous DFO or more likely switching to another chelation regime in the absence of heart failure (see below).

Effects on long term outcome

DFO has been in clinical use since the 1970s and widely used as subcutaneous infusions since about 1980. The most powerful evidence for the effectiveness of DFO and indeed for chelation as a treatment modality is the improving survival and decreased morbidity in patients treated with DFO since this time (Table 6). This benefit is clearly shown in successive cohorts born since this time. Only patients born after 1980 will have started treatment at an early age, and age of starting treatment is a key factor in outcome (Borgna-Pignatti 2004, Brittenham 1993). Regular subcutaneous therapy started before the age of 10 years reduces co-morbidities such as the incidence of hypogonadism (Bronspiegel-Weintrob 1990), as well as other endocrine disturbances, including diabetes mellitus (Borgna-Pignatti 2004, Olivieri 1994, Brittenham 1993). Adherence to therapy has been the main limiting factor to successful outcomes; failure to take treatment at least 5 times a week at adequate doses and subsequent failure to control serum ferritin in the long term leads to increased mortality (Gabutti 1996). Up until 2000, 50% of UK patients still died by age 35 years (Modell 2000), reflecting difficulties with DFO use and other issues such as the variable support patients on chelation therapy received in centres where only small numbers of patients attended. It is important to recognize that toxicity from iron overload is a long-term phenomenon, so the entire chelation history of an individual is important for outcomes, rather than simply the treatment a patient is taking when an event happens.

Table 6.

Decreasing complications in cohorts of Italian patients born after DFO became available. Reproduced with permission from (Borgna-Pignatti 2004).

Standard therapy

• - When to start DFO therapy?
  Provided that treatment is 1) begun within 2-3 years of beginning transfusion therapy, (2) administered regularly (at least 5 times a week) and 3) administered in adequate doses, DFO has a well- established impact on survival and on cardiac and other complications of iron overload described above. In thalassaemia major, this should start before transfusions have deposited enough iron to cause tissue damage. This has not been formally determined, but current practice is to start after the first 10-20 transfusions, or when the ferritin level rises above 1,000 µg/l. If chelation therapy begins before 3 years of age, particularly careful monitoring of growth and bone development is advised, along with reduced dosage.
• - Standard dose and frequency to obtain iron balance
  The standard aim is to balance iron input from transfusions with iron excretion through urine plus faeces. The recommended method is slow subcutaneous infusion over 8-12 hours of a 10% DFO solution, using an infusion pump a minimum of 5 days per week. In countries where pre-filled balloon infusors are available, this has been found to ease the convenience of adhering with DFO chelation. In general, average doses should not exceed 40 mg/kg until growth has ceased. The standard dose is 20-40 mg/kg for children, and up to 50-60 mg/kg for adults, as an 8-12-hour subcutaneous infusion for a minimum of 5-6 nights per week. To achieve negative iron balance in patients with average transfusion requirements, a dose of 50 mg/kg/day at least 5 days a week is required. It is important that patients with high degrees of iron loading, or those at increased risk of cardiac complications receive adequate doses, advice about compliance or consideration of alternative chelator regimens.
• - Use with vitamin C
  Vitamin C increases iron excretion by increasing the availability of chelatable iron, but if given in excessive doses may increase the toxicity of iron. It is recommended not to give more than 2-3 mg/kg/day as supplements, taken at the time of the DFO infusion so that liberated iron is rapidly chelated. Where a patient has just started on DFO and it has been decided to administer vitamin C, the vitamin supplement should not be given until after several weeks of treatment.
• - Dose adjustment to avoid DFO toxicity
  At low ferritin levels, the dose of DFO needs to be reduced and DFO-related toxicities monitored particularly carefully (see below). Dose reductions can be guided using the therapeutic index (= mean daily dose (mg/kg)/SF µg/L) to keep this < 0.025 (Porter, 1989): Although a tool in protecting the patient from excess chelator, this index is not a substitute for careful clinical monitoring. Liver iron concentration has recently been advocated as a more reliable alternative to serum ferritin at low levels of body iron loading (see below).

Rescue therapy

• - Rescue to achieve negative iron balance
  If iron has already accumulated to harmful levels (see monitoring), negative iron balance is necessary. Dose adjustment is critical to the success of chelation therapy; increased frequency, duration and dose when rescue therapy is required, and decreased dosing when body iron is well controlled. Table 7 shows how the dose can be adjusted to achieve negative iron balance, depending on the transfusion rate. At transfusion rates > 0.5 mg/kg/day only about half of patients will be in negative iron balance at doses 35-50 mg/kg, while >50 mg/kg are required to achieve negative iron balance.
• - Rescue to remove cardiac iron
  For patients with mild to moderate myocardial iron (T2* 10-20 ms), increasing the mean daily dose to 50-60 mg/kg/day may be sufficient to improve the T2* provided that adherence to therapy can be achieved. For patients with cardiac iron of 6-10 ms, other chelation regimes have been shown to be effective, such as combination of DFP with DFO or DFX monotherapy (see below). For severe cases of cardiac iron (T2* <6 ms), other regimes need to be considered (see below). For patients with abnormal LVEF, emergency therapy is recommended.
• - Intensive therapy for other reasons
  Prior to pregnancy or bone marrow transplantation, when avoidance of high levels of iron overload is desirable (see Chapters 9 and 12), intensification of therapy may be helpful to minimize the levels of iron overload. The optimal regime has not been studied systematically but may include dose adjustment as described above with attention to adherence through goal setting.

Table 7.

% of responders (% in negative iron balance) by dose and transfusion rate. Adapted from (Cohen 2008).

Emergency therapy

In high risk cases with decreased LVEF, continuous infusion is potentially more beneficial than periodic infusions because it reduces the exposure to toxic free iron (NTBI), which returns to pre-treatment levels within minutes of stopping a continuous intravenous infusion (Porter 1996). The route of administration is not critical, provided that as close to 24-hour exposure to chelation as possible is achieved. Intensification of treatment through continuous, 24-hour intravenous administration of DFO via an implanted intravenous delivery system (e.g. Port-a-cath) (Davis 2000), or subcutaneously (Davis 2004) has been shown to normalise heart function, reverse heart failure, improve myocardial T2* (Porter 2013b, Anderson 2004) and lead to long-term survival, provided treatment is maintained. Some studies have included cases where for operational reasons, intensification was undertaken without continuous infusion. Continuous infusion is usually given through an indwelling line for long-term management. For emergency management before a central line can be inserted, DFO can be given through a peripheral vein, provided it is diluted in at least 100 mls of saline to avoid damage to the veins where the drug is infused. A dose of at least 50 mg/kg/day and not exceeding 60 mg/kg/day is recommended as a 24-hour infusion (Davis 2004, Davis 2000). Higher doses have been used by some clinicians, however, DFO is not licensed at these doses and the risk of retinopathy increases. Addition of vitamin C is recommended only when acute heart dysfunction has settled, which usually occurs by three months of continuous treatment (Anderson 2004). As ferritin falls, the dose but preferably not the duration of treatment can be reduced - in line with the therapeutic index (see above).

The question of whether to add DFP to intensified DFO needs to be considered. This is partly because DFP at high doses (90-100 mg/kg) was found to increase the T2* more than conventional s.c. DFO 5 days a week (Pennell 2006b) and because combined DFP + DFO has also been found to improve T2* more rapidly than conventional doses of DFO (Tanner 2007). However, these patients have baseline LVEF in the normal range and were not in heart failure but showed greater LVEF increases with the DFP containing regimes. Furthermore, these studies compared conventional intermittent non-intensified DFO with the DFP containing regimes; such low DFO doses should not be recommended for patients in heart failure. The only randomised study to examine the effect of additional DFP to intensified DFO found no difference between the two study arms with or without DFP, either with respect to LVEF or to improvements in T2* (Porter 2013b). Nevertheless, this study also showed no major extra toxicities in the study arm containing DFP, so the addition of DFP to intensified DFO would seem a reasonable course of action for patients in heart failure, provided that patient can tolerate oral administration of DFP.

Effects on ferritin

Randomised trials comparing the effects of DFP on serum ferritin at baseline and at follow-up have been reported from the 1990s (Pennell 2006b, Ha 2006, Gomber 2004, Maggio 2002, Olivieri 1997). Pooled analysis shows a statistically significant decrease in serum ferritin at six months in favour of DFO, with no difference between the two drugs at 12 months (Pennell 2006b, Gomber 2004). There are numerous non-randomised cohort studies demonstrating a lowering of serum ferritin at doses of 75 mg/kg/day administered in three doses. The effect on SF at this dose appears greater at higher baseline ferritin values. In these studies significant decreases in serum ferritin are seen in patients with baseline values above 2,500 µg/L (Viprakasit 2013, Olivieri 1995, Al-Refaie 1992, Agarwal 1992) but not with values below 2,500 µg/L (Cohen 2000, Hoffbrand 1998, Olivieri 1995). In a recent study from Thailand, only 45% of paediatric thalassaemia patients (age > 2 years) had significant reduction of serum ferritin after 1 year at doses of over 79 mg/kg/day (Viprakasit 2013). In this study, baseline SF was the major factor that predicted clinical efficacy; patients with baseline SF>3,500 µg /L had the most significant fall of SF at 1 year. The FDA licensing agreement in 2011 concluded that “data from a total of 236 patients were analyzed, of the 224 patients with thalassemia who received DFP monotherapy and were eligible for serum ferritin analysis…the endpoint of at least a 20% reduction in serum ferritin was met in 50% (of 236 subjects), with a 95% confidence interval of 43% to 57%”.

Effects on liver iron

Change in LIC from baseline after various periods of treatment with DFP has been compared with DFO in randomised studies (El-Beshlawy 2008, Ha 2006, Pennell 2006a, Maggio 2002, Olivieri 1998) and also with combination of DFP plus DFO (Aydinok 2007). One study showed initial LIC decreases at 1 year but increases in LIC at 33 months of 5 mg/g dry wt with DFP (n=18) and 1 mg/g dry wt with DFO (n=18) (Olivieri 1998). In another study an average decrease in LIC at 30 months was reported with both DFP (n=21) and DFO (n=15) (Maggio 2002). A greater 1 year decrease in LIC with DFO than DFP monotherapy was reported in several studies (Aydinok 2007, Pennell 2006a). A decrease of 0.93 mg/g dry wt with DFP (n=27) and 1.54 mg/g dry wt with DFO (n=30) (Pennell 2006) was observed at 1 year. Another study reported initial decreases in LIC at six months with both DFP and DFO, but LIC had increased by the end of the trial (Ha 2006), consistent with earlier observations (Olivieri 1998). In a randomised 1 year comparison of DFP with DFP + DFO, there was no decrease in LIC with DFP monotherapy but a decrease with combination therapy or in the DFO comparison group (Aydinok 2007). In a non-randomised prospective study using DFP, LIC increased with DFP by 28% at two years and by 68% at three years (Fisher 2003). In a recent study of paediatric patients, decrease in LIC was seen in those patients who showed a clinical response by reduction of serum ferritin and in those with a higher baseline LIC (Viprakasit 2013). In observational studies where only single biopsies were performed after several years of DFP treatment, LIC was found to be above 15 mg/g dry wt in variable proportions of patients, ranging between 11% (Del Vecchio 2002), 18% (Tondury 1998) and 58% (Hoffbrand 1998). Overall negative iron balance (decrease in LIC) with standard transfusion rates using DFP monotherapy is achieved in only about 1/3 of patients receiving 75 mg/kg (Fischer 1998).

Effects on myocardial iron

The effect of DFP monotherapy on myocardial iron has been reported in randomised studies. One compared high dose DFP (92 mg/kg/day) with s.c. DFO 5-7 days a week in patients with mild to moderate myocardial iron (mT2* 8-20 ms). The actual dose prescribed for DFO was 43 mg/kg for 5.7 days/week (or a mean daily dose of 35 mg/kg/day). The increase in mT2* from 13 ms to 16.5 ms in the DFP group was greater than that seen in the DFO group, where an increase from 13.3 to 14.4 ms at 1 year was observed (Pennell 2006b). In another 1 year randomised study of DFP and DFO, no change in heart iron estimated by MRI (signal intensity ratio) was reported for either drug. Lower doses of DFP (75 mg/kg/day) were used in this study (Maggio 2002). In a retrospective study, higher mT2* values were seen using a multislice technique and with higher global systolic ventricular function, in patients with DFP monotherapy (n=42) than those with DFO (n=89) or DFX (n=24) monotherapies, although the mean values were in the normal range in each monotherapy category (Pepe 2011).

Effects on heart function

The effects of DFP on heart function have been documented in patients with normal baseline function (Pennell 2006b, Maggio 2002) but not in patients with baseline LVEF below the normal reference range. In a one-year randomised study of patients with normal LVEF, DFP given at high doses (92 mg/kg) increased LVEF (Pennell 2006b). In another 1 year randomised study, no difference in LVEF or other measures of LV function were seen with either DFP at 75 mg/kg/day or DFO (Maggio 2002). In a three year retrospective reanalysis of patients in the one year prospective study, follow up data showed that DFP monotherapy was associated with significant increase in LVEF in patients with LVEF in the normal range at baseline (Maggio 2012). Another retrospective study of 168 patients with thalassaemia major and baseline mean LVEF within the normal range were followed for at least 5 years while receiving monotherapy with DFO or DFP. LVEF increased in both groups but was higher in the DFP group at 3 years. However the subgroup of patients with LVEF <55% at baseline was greater in the DFO than in the DFP group (Filosa 2013).

Compliance with DFP

One study comparing compliance with DFP and DFO found rates of 95% and 72% respectively (Olivieri, 1990), while another reported rates of 94% and 93%, respectively (Pennell 2006b). The similar rate of compliance with DFP has been observed in other populations (Viprakasit 2013). As with other oral chelators, two important points should be taken into consideration: (i) compliance with any treatment tends to be higher in in the context of clinical studies than in routine use, and (ii) although compliance with oral treatment is expected to be better, the importance of constant supervision and patient support as provided when administering DFO, should not be overlooked.

Unwanted effects with DFP

The unwanted effects of DFP and their monitoring and management are described in Appendix 2.

Recommended treatment regimens with DFP

According to the FDA, Ferriprox ^® ‘’is indicated for the treatment of patients with transfusional iron overload due to thalassemia syndromes when current chelation therapy is inadequate’’ (FDA 2011). FDA approval is ‘based on a reduction in serum ferritin levels’. The European licensing Agency (EMEA) states ‘Ferriprox is indicated for the treatment of iron overload in patients with thalassaemia major when DFO therapy is contraindicated or inadequate’. In Thailand and many Asian countries, DFP was registered for similar indications and is licensed for use from the age of 6.

Standard dosing and frequency

The daily dose of DFP that has been evaluated most thoroughly is 75 mg/kg/day, given in three doses. In the EU, the drug is licensed for doses up to 100 mg/kg/day but formal safety studies of this dose are limited. The standard dose of 75 mg/kg/day administered in three separate doses is therefore recommended. The drug’s labeling includes charts stating how many tablets and half tablets to use per dose for patient weights ranging from 20 to 90 kg. Each 500 mg tablet is scored to facilitate tablet splitting. An oral solution is also available for paediatric use.

Dose escalation with DFP

Adjustments may be made on the basis of the patient’s response but should never exceed 33 mg three times daily. Doses of 100 mg/kg/day have been given in at least one prospective study (Pennell, 2006), with no increase in reported side-effects. The relation of dose to iron balance or serum ferritin has not been reported in a single study. High dose monotherapy with DFP has not yet been prospectively evaluated for safety and effectiveness for patients with abnormal heart function, and thus combination therapy with DFP and DFO (see below) or intensive therapy with DFO as a 24-hour infusion should be recommended for this group of patients.

Age of commencement

There is less experience on the safety and efficacy of DFP in children under 6 years of age than in adults. A recent open label prospective study examined efficacy and tolerability in 73 pediatric patients, age range 3-19 years (Viprakasit 2013), as well as a similar study involving 100 children of 1-10 years old who received the liquid formulation of DFP found no specific tolerability issues that have not been previously reported in adults.

Use of vitamin C

The effect of vitamin C on iron excretion with DFP is not clear and is thus not recommended.

Safety monitoring, precautions and interactions

These are summarised in Appendix 1 and described in Appendix 2.

Chemistry and Pharmacology

Deferasirox is an orally absorbed tridentate iron chelator, with two molecules binding each iron atom. The chemical properties and pharmacology are summarized in Table 5. The tablet is dispersed (not dissolved) in water or apple juice using a non-metallic stirrer and consumed as a drink once daily, preferably before a meal. The drug is rapidly absorbed, reaching peak concentrations of 80µM at 20 mg/kg and the long half-life of this iron-free drug allows trough concentrations of about 20 µM, providing 24hr protection from plasma labile iron (Nisbet-Brown 2003, Galanello 2003), with about 90% in the free drug form and 10% as iron complexes (Waldmeier 2010). The lipid solubility allows entry into cells, including cardiomyocytes. The majority of the drug is excreted in faeces, and metabolism is mainly to an acyly-glucuronide that retains its ability to bind iron (Waldmeier 2010). Metabolic iron balance studies show iron to be excreted almost entirely in the faeces, with less than 0.1% of the drug eliminated in urine (Nisbet-Brown 2003). The main pathway of DFX metabolism is via glucuronidation to the acyl glucuronide and the 2-O-glucuronide metabolites. Oxidative metabolism by cytochrome 450 enzymes is minor (10% of the dose) (Waldmeier 2010). The efficiency of chelation is 28% over a wide range of doses and levels of iron loading.

Dose effect on serum ferritin

A dose-dependent effect on serum ferritin has been observed in several studies (Porter 2008, Cappellini 2006, Piga 2006). A prospective randomised study comparing the effects of DFX in 296 thalassaemia major patients with DFO in 290 patients, found that 20 mg/kg daily stabilized serum ferritin close to 2,000µg/L and at 30 mg/kg, serum ferritin was reduced with an average fall of 1,249 µg/L over one year (Cappellini 2006). Longer-term analysis of ferritin trends show that the proportion of patients with ferritin values <1,000 µg/L and less than 2,500 µg/L is decreasing progressively with time. At 4-5 years follow up in 371 patients, median SF had fallen to < 1500 µg/L (Cappellini 2011) and the increase in mean dose from an initial value of < 20 mg/kg to 25 mg/kg was associated with a significant fall in serum ferritin. Overall, 73% of patients attained serum ferritin levels ≤2500 µg/L and 41% of patients achieved serum ferritin levels of ≤1000 µg/L, compared with 64% and 12% at baseline respectively. A large-scale prospective study (EPIC) has examined the interaction between dose and SF response in large scale studies involving 1,744 transfusion-dependent anaemias, including 1,115 with TM (Cappellini 2010). The initial dose of deferasirox was 20 mg/kg/day for patients receiving 2-4 packed red blood cell units/month, and 10 or 30 mg/kg/day for patients receiving less or more frequent transfusions, respectively. Dose adjustment were made on the basis of ferritin trends at 3 monthly intervals. A significant though modest overall fall in ferritin was seen at 1 year. In a recent substudy, the largest SF decrease of -1,496 µg/L/year was noted in patients with the highest baseline SF values (baseline median SF 6,230 µg/L) (Porter 2013a). These patients were treated with DFX at high dosage (35-40 mg/kg/day), which are therefore doses now recommended for heavily iron overloaded patients.

Dose effect on liver iron and iron balance

Metabolic balance studies showed that excretion averaged 0.13, 0.34 and 0.56 mg/kg/d at DFX doses of 10, 20 and 40 mg/kg/d respectively, predicting equilibrium or negative iron balance at daily doses of 20 mg and above (Nisbet-Brown 2003). In a longer term randomised prospective study in 586 thalassaemia patients aged 2 to 53 years (with half of patients <16 years old), iron balance with DFX (n=290) assessed by serial LIC determination was achieved at 20 mg/kg/day, with mean LIC remaining stable over one year (Cappellini 2006). Negative iron balance was achieved at 30 mg/kg/day, with a mean LIC decrease of 8.9 mg/g dry wt (equivalent to a decrease in body iron of 94 mg/kg body weight) over one year. These are average trends and a closer analysis shows that the blood transfusion rate influences the response to treatment (Cohen 2008) (Table 8). This shows that negative iron balance over 1 year (response rate) is increased as doses increase, and that the response rate is less at high transfusion rates, who therefore required higher doses.

Table 8.

% of responders (% in negative iron balance) by dose and transfusion rate. Adapted from (Cohen 2008).

At 4-5 years of follow up, the percentage of patients with LIC values <7 mg/g dry wt by biopsy increased from 22% at baseline to 44% (Cappellini 2011). A more moderate reduction in LIC occurred in children under six years old, despite the administration of an average dose of 21.9 mg/kg in this subgroup. However, these patients had the highest mean transfusional iron intake. In a recent liver MRI analysis of 374 patients enrolled on the EPIC study (Porter 2013a), response to DFX was analyzed according to baseline levels of iron overload. In patients with a high baseline LIC of 27.5 mg/g dry wt, LIC decreased by 6.9 mg/g dry wt at one year at doses of 25-35 mg/kg/day. In patients with LIC of 32 mg/g dry wt the decrease was 7.3 mg/g dry wt and 35-40 mg/kg/day, respectively. Thus, provided adequate doses are given, there is a good response to DFX across the full range of baseline LIC values (Porter 2013a).

Iron balance and safety in children

DFX was the first chelator to be formally assessed in children as young as 2 years old. Approximately 50% of patients in 5 clinical studies that included 703 patients were children aged <16 years. The drug appears to be tolerated in children as well as in adults. Importantly, no adverse effects on growth or skeletal development were observed at a dose of 10 or 20 mg/kg /day (Piga 1988). In another observational study of chelation-naive transfusion-dependent children (aged < 5 years) with SF > 1000 µg/L at baseline, DFX or DFO was prescribed to maintain serum ferritin levels between 500 and 1000 µg/L. With a median follow up of 2.3 years for DFX (n = 71) and 2.8 years for DFO (n = 40), DFX was shown to be well tolerated and at least as effective as DFO in maintaining safe serum ferritin levels and normal growth progression (Aydinok 2012b).

Effect on myocardial T2*

Improvement in mT2* was first reported in a retrospective analysis of effects on myocardial T2* after 1 and 2 years (Porter 2010, Porter 2005a). Prospective data demonstrated the efficacy of DFX in improving myocardial T2*over a range of mT2* from 5-20 ms, with 41% having severe myocardial iron loading <10 ms at baseline (Pennell 2010). In a prospective trial, 114 patients with high mean baseline LIC (mean 28 mg/g dry wt) were treated with DFX for up to 3 years (Pennell 2012), receiving mean actual doses of 33, 35, and 34 mg/kg/day during the 1st, 2nd and 3rd years, respectively. Higher mean doses of 37 mg/kg per day were received by patients with baseline T2* between 5 and <10 ms, compared with those between 10 and 20 ms (32 mg/kg per day). Of the 114 patients initially enrolled, 101 continued into the 2nd year, 86 completed two years of treatment and 71 entered into a third year. There was year by year significant improvements in mT2*; from 12.0 ms at baseline to 17.1 ms at 3 yrs, corresponding to a decrease in cardiac iron concentration (from 2.43 mg/g dry wt at baseline, to 1.80 mg/g dry wt. After three years, 68 % of patients with baseline T2* between 10 and <20 ms benefited from normalization of T2*, and 50% of patients with baseline T2*>5 to <10 ms at baseline improved to 10 to <20 ms. There was no significant variation in left ventricular ejection fraction over the three years and no deaths occurred. Tolerability was similar to other DFX studies in TM over the doses up to 40 mg/kg/day.

In a 1 year randomised prospective study (CORDELIA) 197 patients with T2* of 6-20 ms and no signs of cardiac dysfunction were randomised to DFX (target dose 40 mg/kg/day) or subcutaneous DFO treatment (50-60 mg/kg/day for 5-7 days/week) (Pennell 2014). Baseline LIC was high in both DFX (mean 29.8 mg/g dry wt) and DFO treated patients (30.3 mg/g dry wt), with 73% of patients having baseline LIC >15 mg/g dry wt. The geometric mean (Gmean) myocardial T2* improved with DFX from 11.2 ms at baseline to 12.6 ms at 1 year (Gmeans ratio 1.12) and with DFO (11.6 ms to 12.3 ms, Gmeans ratio 1.07). This study established non-inferiority of DFX vs. DFO for cardiac iron removal in this patient population. LVEF remained stable in both arms and the frequency of drug-related adverse events was comparable between DFX (35.4%) and DFO (30.8%). Taken together, these studies show that DFX is an effective treatment for patients with increased heart iron with mT2* >5-20 ms. It also demonstrates response in patients with high levels of baseline mT2* (5-10 ms), as well as those with high levels of baseline LIC or SF. As with other chelation regimes, high levels of baseline heart iron (<10 ms) will typically take several years to clear, but the risk of developing heart failure during this time appears very low (see below), provided treatment is monitored.

Effects on heart function

In the above studies, even though mT2* values at baseline were as low at 5-6 ms and the proportion of patients with mT2* <10 ms was significant (17.2-33%), LVEF remained stable, and there were neither deaths nor episodes of symptomatic heart failure observed. Only one case of atrial fibrillation and one case of cardiomyopathy were reported. According to risk analysis of heart failure in TM from other cohorts, the risk of developing cardiac failure was expected to be substantial, with a relative risk 160 fold higher for patients with T2* < 10 ms (Kirk 2009a). The stability of LVEF and the absence of heart failure in this otherwise high risk group of patients suggests that DFX renders effective prophylaxis for heart failure, even in patients with T2* values of 5-10 ms. This may be related to the 24hour ‘protection time’ against labile iron that results from the long plasma half-life of DFX (Daar 2009). Deferasirox has not been evaluated in formal trials for patients with symptomatic heart failure or LVEF < 56%, therefore at this time other chelation options are recommended for such patients.

Evidence of long term survival benefits of DFX

More than 5,900 patients have been enrolled in prospective trials but these, with some exceptions, have typically been designed for short term evaluation. Up to 5 years of follow up have now been reported in one prospective clinical trial from the initial registration studies, which provides useful information about risk and benefit with this treatment (Cappellini 2011). Other prospective data of patients with myocardial T2* of 5-20 ms and high levels of liver iron but without complications, provides further insight into comorbidities and mortality in high-risk subjects. The stability of left ventricular function, lack of progression to heart failure and absence of any deaths are notable features of the prospective 3-year EPIC and 1-2 year CORDELIA cardiac studies, despite including patients at high risk of cardiac decompensation, with mT2* levels as low as 5 ms (Pennell 2012) or 6 ms (Pennell 2014).

Convenience and impact on quality of life

Convenience and quality of life on DFX, as with other oral chelation regimes, are expected to impact on adherence and hence survival. This is likely to have a greater impact outside formal clinical studies, where adherence is generally better than in routine clinical use. Studies comparing satisfaction and convenience of DFX with DFO in thalassaemia major show a significant and sustained preference for DFX (Cappellini 2007). In a randomised comparison, total withdrawals in DFX-treated patients was 6% at one year, compared with 4% with DFO (Cappellini 2006). This compares with a dropout rate of 15% at one year with DFP, although these are not matched populations (Cohen 2000). In the large scale EPIC study, patients reported improved quality of life (estimated by SF36 scores) and greater adherence to chelation therapy compared with baseline before starting DFX (Porter 2012).

Recommended standard dosing

Deferasirox is taken orally as a suspension in water once daily, and preferably before a meal. A starting dose of 20 mg/kg is recommended for thalassaemia major patients who have received 10-20 transfusion episodes and currently receive standard transfusion at rates of 0.3-0.5 mg of iron/kg/day. In those patients in whom there is a higher rate of iron intake from transfusion (>0.5 mg/kg/day), or in patients with pre-existing high levels of iron loading where a decrease in iron loading is clinically desirable, 30 mg/kg/day is recommended. For patients with a low rate of iron loading (<0.3 mg/kg/day), a dose of 10-15 mg/kg may be sufficient to control iron loading.

Age of commencement

The labeling for age of commencement differs in countries that follow US licensing from those that adhere to EU licensing (Appendix 1). However, based on prospectively randomised studies of DFX in children as young as two years of age, some recommendations can be made. A fall in LIC has been seen across all age groups analysed, with no age-related adverse effects. In particular, no adverse effects on growth, sexual development or bones have been observed (Piga 2006). Deferasirox also appears to be palatable to children at this young age. On the basis of present knowledge, the criteria for starting treatment (ferritin level, age, number of transfusions) are similar to those of DFO. However, a target of 500-1000 µg/L appears to be achievable with DFX without additional toxicity issues, provided that doses are adjusted downwards as SF values fall towards 500 µg/L.

Rescue therapy to achieve negative iron balance

When body iron has accumulated to high levels (see monitoring), negative iron balance is required. The proportion of patients in negative iron balance at a given dose is partially dependent on the rate of iron loading (see above). Doses of up to 40 mg/kg/day are recommended for patients with LIC or SF values and are now licensed at this dose (Porter 2013a). Dividing the dose as a twice daily dose has been used in some patients who fail to achieve negative iron balance, despite these higher doses (Pongtanakul 2013). Some patients have taken DFX after rather than before food, with apparently improved efficacy. This is consistent with the known effects of food on GI absorption (Galanello 2008).

Rescue therapy for patients with mild to moderate myocardial iron (5-20 ms)

On the basis of prospective studies these patients can be successfully treated with DFX, resulting in preservation and stabilisation of LV function. Doses of up to 40 mg/kg have been used and are advisable in patients with very high levels of liver iron or serum ferritin.

Rescue therapy for patients with severe myocardial iron (< 6 ms)

Prospective clinical trials with DFX monotherapy have been confined to patients with mT2* values ≥6 ms. For patients with mT2* <6 ms, other alternative chelation regimes are recommended.

Emergency therapy for patients with reduced LVEF or symptomatic heart failure

DFX has not been formally evaluated in prospective trials for such patients and is therefore not recommended.

Other indications and contraindications

DFX is contraindicated in patients with renal failure or significant renal dysfunction (see below). Caution is recommended for patients with advanced liver disease and hepatic decompensation.

Unwanted effects with DFX

Unwanted effects of DFX and their monitoring and management are described in Appendix 2.

Concept and pharmacology of combination therapies

The term ‘combination therapy’ has been used to cover a variety of approaches to improve outcomes if monotherapy proves inadequate. In principle, two chelators can be given at the same time (simultaneously), or one after the other (sequentially). True combination, where two chelators are present in the blood at the same time, has been used relatively rarely compared with sequential regimes. Some investigators have used the term ‘alternating therapy’ to describe the use of two drugs administered on alternate days, reserving the term ‘sequential therapy’ for when DFO is given at night and DFP during the day. In practice, regimes may involve both a component of ‘sequential’ and ‘alternating’ therapy such as when DFO is given three times a week (alternate nights) and DFP every day. Most commonly used regimes have tended to give DFP daily at standard doses, combined with varying frequency and dosing of DFO. More recently, combinations of DFX with DFO, or DFX with DFP have been evaluated. In this instances, both drugs may be present in plasma or intracellularly for at least part of the time owing to the half-life of DFX and its extended time in the plasma - for up to 24hr.

The pharmacology and mechanisms of action in combining chelators is dependent on whether the drugs are present in cells or plasma at the same time. By giving DFO at night and DFP by day, 24-hrs of exposure to iron chelation can be achieved (similar exposure to that achieved with 24-hrs desferrioxamine infusion, or once daily deferasirox). This has the theoretical advantage of 24hr protection from labile (redox active) iron (Cabantchik 2005). If the drugs are given at the same time (simultaneously), they may interact in a process that involves the ‘shuttling’ of iron, which may lead to additional chelation of iron from cells or plasma NTBI (Evans 2010) and so improved efficiency of iron chelation. On the other hand, there is also the possibility of chelation from metalloenzymes, leading to increased drug-related toxicity, but this has not been an issue clinically. The use of DFX, which is present in plasma 24hrs/day, together with DFO by intermittent infusion provides 24hr chelation, with decreases in LPI and NTBI (Lal 2013). Simultaneous exposure to two chelators may also result in synergistic removal of cellular iron. This has been demonstrated in cell culture with combinations of all three chelators (Vlachodimitropoulou 2013).

Evidence of efficacy of combined regimes

• - Effects of sequential use on serum ferritin
  One study (Mourad 2003) found the decrease in SF achieved using five days of DFO monotherapy (n=11) to be similar to that achieved with two nights of DFO plus seven days of DFP at 75 mg/kg (n=14). Another randomised study involving 30 patients and three different treatments (Gomber 2004) found that the decrease in SF was greatest with five nights of DFO, albeit not significantly different from that achieved with a combined treatment of DFO two nights a week, plus DFP seven days a week. A further randomised study involving 60 patients receiving ‘alternating’ therapy (Galanello 2006) found no difference in SF in patients randomised to combined treatment (two days DFO at 33 mg/kg + seven days DFP at 75 mg/kg), or to DFO five nights a week at 33 mg/kg. In another randomised study from Turkey (Aydinok 2007), SF decreased more with combination therapy than DFP monotherapy but similarly to DFO monotherapy. In a randomized study of 65 patients (Tanner 2007), serum ferritin was decreased more by combined treatment (DFO five days a week plus DFP seven days a week) than with standard DFO monotherapy (40 mg/kg five times a week). A 5-year follow up randomized clinical trial (Maggio 2009) also showed a greater SF reduction with sequential DFP (75 mg/kg/day-4 days/week)-DFO (50 mg/kg/day-3days/week) compared with DFP alone on reducing serum ferritin, with comparable adverse effects and cost. However, survival trend of both groups was not significantly different. Taken together, these studies show that serum ferritin can be controlled with a relatively low frequency of DFO given twice a week when combined with DFP at standard doses (75 mg/kg/day).
• - Effects of sequential use on liver iron
  When an alternating DFP + DFO regime was compared to DFO monotherapy, a baseline LIC of <7 mg/g dry wt was maintained on average in both arms of the study (Galanello 2006). A different prospective randomised study found LIC reduction in either DFO monotherapy or DFP + DFO (2 times weekly) combination therapy, but not with DFP monotherapy at 75 mg/kg/day (Aydinok 2007). In another study comparing LIC changes using DFP + DFO or DFO monotherapy (5 times a week), a greater improvement in liver T2* (as a surrogate measure of LIC) was seen with combination therapy than with DFO alone (Tanner 2007).
  In another randomised study in patients with heart failure who received DFO with or without DFP, there was decrease in LIC and ferritin in the combination arm but not with monotherapy (Porter 2013b). These studies did not give definitive results with respect to the comparative LIC effects of combination therapy relative to other therapies, although they do support combination therapy being more effective than DFP monotherapy. The relative efficacy of combination compared with DFO monotherapy most likely depends on the dosing and frequency of DFO used in the different regimes.
• - Effects on heart function and T2*
  In a randomised controlled study of 65 patients without heart failure but with moderate heart iron loading (T2* 8-20 ms) and LVEF in the normal reference range, changes in myocardial T2* with combined DFP at 75 mg/kg seven days a week plus DFO five days a week were compared with patients on standard DFO treatment five times a week (Tanner 2007). The study showed that LVEF increased within the normal range by approximately 2.5% in the combination arm and 0.5% in the DFO monotherapy arm. Myocardial T2* improvements were seen with both treatments, but were greater in the combination arm. For high risk patients with decreased LVEF or with symptomatic heart disease, a prospective randomised study showed significant improvement in LVEF and T2* in patients receiving either DFO intensification or DFO intensification plus DFP (Porter 2013b). No statistical difference between the two study arms was found with respect to these variables, although larger study numbers would be required to show a 5% difference in LVEF response. Observational studies have also reported changes in heart function with combined treatment. In 79 patients with SF>3000 µg/L treated with a variable DFO regimen plus DFP at 75 mg/kg seven days a week for 12 to 57 months, there was an increase in LVEF by echocardiography (Origa 2005). In an observational study of 42 patients with sequential use of treatment over three to four years (DFP 75 mg/kg/day plus DFO two to six days a week), the LV shortening fraction improved (Kattamis 2006). In another observational multicentre study, changes in mT2* and LVEF were examined in patients who were receiving combined DFO + DFP (N=51), DFP (N=39) or DFO (N=74) monotherapy (Pepe 2013). The proportion of patients that maintained a normal heart T2* value was comparable between DFP and DFO when compared with both monotherapy groups. Increase in mT2* was greater in patients on DFO + DFP, or DFP than with DFO monotherapy, but did not differ between DFO + DFP and DFP, with combination therapy not showing an additional effect on heart function over DFP alone. Monotherapy with DFX is usually effective at improving T2* across a full range of LIC levels (Pennell 2012). The effects of DFX on T2* has not been compared directly with DFP either alone, or in combination with DFO. If the imperative is to have improvement as rapidly as possible (for example in preparation for pregnancy or prior to bone marrow transplant), then high dose DFP monotherapy (>90 mg/kg) or regular doses of DFP in combination with standard DFO 5 days a week should be used, as combination of DFP and DFO may result in more rapid improvement in cardiac iron load and increases in LVEF, which has not been observed with other chelation regimes.
• - Long term effects on survival
  Several retrospective analyses have reported a benefit in survival compared with DFO alone. In one such analysis, of 544 patients with β-thalassemia from Cyprus treated between 1980 and 2004, 304 switched to combination chelation therapy from 1999 as part of their regular treatment. The authors reported a worsening survival in Cyprus up until 2000, followed by an improved subsequent survival which they attributed to switching to combination therapy (Telfer 2009, Telfer 2006). 75 patients came off combination therapy because of agranulocytosis (5%), recurrent neutropenia (2.9%), gastrointestinal disturbances (5.6%), arthralgia (1.6%), allergic reactions (0.7%), weight gain (0.7%), increased liver enzymes (0.3%), non-adherence to DFO (3.3%), pregnancy (2.6%), and other reasons (2%). Some authors have achieved very low serum ferritin values using a ‘flexible’ approach to combination therapy without any reported side effects, although the reasons for giving combination therapy were not clear (Farmaki 2010). Controlled trials demonstrating improvement in disease-related symptoms, organ function, or increased survival are, however, lacking. Meta-analysis reviewing multiple trials found no statistically significant variations in heart T2* signal during associated or sequential versus mono-therapy treatment (p=0.46 and p=0.14, respectively) (Maggio 2011). This analysis did find “improved ejection fraction during combination associated or sequential (combination) versus monotherapy treatment (p=0.01 and p=0.00001, respectively)”. However ‘’these findings do not support any specific chelation treatment. The literature shows risks of bias, and additional larger and longer trials are needed’’ (Maggio 2011).
• - Safety of combined DFO + DFP treatment
  Formal safety data on combined treatment are limited. The side effects described above are largely consistent with the known effects of the individual chelating agents, with the possible exception of cerebella syndrome in a single case. Tolerability of simultaneous combinations may differ from sequential use, but this has not been formally studied.
• - Conclusions and possible treatment regimens
  Combinations of these two drugs are useful, especially when various monotherapy regimes have failed to control either liver iron or cardiac iron. This is most commonly a consequence of pharmacological differences with DFP, or for reasons of poor adherence to DFO. In general, if a patient is not doing well with DFP monotherapy, combined treatment offers an additional option to improve iron balance. For patients not doing well with DFO monotherapy for reasons of compliance, and where dose intensification has failed, combined treatment has been used as a way of decreasing the frequency at which DFO is needed to maintain SF and iron balance. For patients with very high levels of heart iron or cardiac dysfunction without frank heart failure, 24-hour treatment with DFO and daily therapy with DFP should be strongly considered.

Standard therapy for obtaining iron balance

The licensing of individual chelators, specified in the country where the treatment is prescribed should act as an initial guide on when to start the therapy and at what dose (see Appendix 1). Standard first line doses have been discussed in detail above, and depend in part on the rate of transfusional iron loading. Starting chelation before overload has built up, or irreversible damage has occurred is critical to success. With DFO, chelation was often withheld until the SF had reached 1000 µg/L because of fears toxicity would have on growth, ears and eyes at low levels of body iron. It may be that with new chelator regimes that chelation can be started earlier, however, information about this is limited at present. In practice, the exact timing of starting chelation is currently constrained to some extent by the licensing of the compound by regulatory authorities, which differs somewhat between countries. If a patient is failing on first line therapy, dose adjustment and attention to adherence (practical as well as psychological support) are the next steps. If this fails then regime adjustment can be considered, depending on the circumstances - some of which described below.

Iron load too high or is increasing – rescue therapy to achieve negative iron balance

If body iron load builds up because of a delayed in starting chelation therapy, inadequate dosing, and poor adherence or because of poor response to an individual monotherapy, rescue therapy is required by one or more of the following:

1. Increasing the dose of chelation a, b, c
2. Increasing frequency of the chelator (improving adherence d, or increasing prescription advice)
3. Switching chelator regimen
4. Rotating e or combining f chelators
   1. DFO monotherapy is effective at producing negative iron balance if it is given in sufficient doses and sufficient frequency, but adherence is often a problem.
   2. Dose escalation of DFX is effective at producing negative iron balance (see Table 7). Doses >35 mg/kg and up to 40 mg/kg are effective in patients with high LIF or SF.
   3. DFP monotherapy is likely to achieve iron balance at 75 mg/kg in only about one third of patients, with average transfusion rates. DFO is often added to this.
   4. If adherence is the major reason for a regime not working, every effort should be made by the health team to support the patient and their family in achieving better adherence.
   5. Rotation of individual monotherapies (sequential chelation) can be helpful in managing individual patients, often for reasons of adherence as much as for specific complications.
   6. True ‘combination therapy’ (where two chelators are combined with some degree of overlap pharmacologically) is widely practiced, although not specifically licensed. This has been shown to be useful in individual patients when monotherapy is inadequate, either to control iron balance or to control iron distribution, particularly in the heart.

Mild increase in cardiac iron (T2* 10-20 ms) - rescue therapy to remove heart iron

DFO, DFP and DFX monotherapy are all effective at decreasing heart iron, but need to be given without interruption for optimal effects and at adequate doses. The immediate risk of heart failure is low, provided that the patient remains on chelation therapy without interruption (Kirk 2009b). Regular daily monotherapy at optimal doses (often an increase from current dose or frequency) will usually improve heart iron but normalisation of heart iron can take several years of consistent therapy and sequential monitoring of T2* is required. Monotherapy with DFX is usually effective and recent work shows that DFX is effective at improving T2* across a full range on LIC concentrations (Pennell 2014, Pennell 2012). If the imperative is to do this as rapidly as possible (for example in preparation for pregnancy or prior to bone marrow transplant), then high dose DFP monotherapy (>90 mg/kg) or regular doses of DFP in combination with standard DFO 5 days a week should be used, rather than DFO alone when given s.c. 5 days a week. DFX has not been compared directly with DFP either alone or in combination with DFO. If there is no trend of improvement in T2* with DFO, DFP or DFX monotherapy, then combined DFP and DFO should be considered.

Cardiac T2* < 10 ms - rescue therapy for heart iron

The risk of developing heart failure increases with lower cardiac T2*, especially when values drop below 10 ms (i.e. higher heart iron). However, if continuous chelation therapy is given, heart failure may be prevented even before the T2* is corrected. This was has been shown for continuous 24h DFO, with high dose DFX in a population where T2* was 6-10 ms, and in patients treated with different combination regimes (DFO+DFP, DFO+DFX). Patients with T2* ≤ 6 ms, are a very high risk group for developing heart failure. This group has not been evaluated extensively with interventional studies (expect people with heart failure). There is some experience of treating these patients with DFO + DFP, but randomised trials did not include patients with T2* values < 8 ms (Tanner 2007). In the absence of formal comparisons with other regimes, the combination of DFO (given as often and as continuously as possible) with DFP at standard doses is recommended. DFX monotherapy at doses >30 mg/kg/day has also been shown to be effective for patients with T2* >5 ms and normal heart function. If patients also have high levels of body iron as well as heart iron, it is important that the regime also reduces total body iron.

Patients with heart failure - reverse heart failure

If chelation therapy is taken regularly clinical heart failure is now rare. Reversal of heart failure requires continuous DFO therapy and can occur within a few weeks of starting treatment. This will not succeed in all cases, but if started early in the development of heart failure, is usually effective. The addition of DFP in these circumstances may be beneficial, although a small randomised comparison did not show a difference with our without DFP (Porter 2013b). Once reversal of heart failure has been demonstrated clinically and with myocardial MRI or echocardiography, continuation of the same therapy is recommended until the cardiac T2* improves to T2* above 8 ms, depending on the starting T2* this may take as long as a year. The key to success is the timely introduction of intensification and the maintenance of intensive treatment after the heart failure has been corrected.

Downward adjustment of chelator dose if body iron falls rapidly or reaches low levels

An increasingly common problem for patients who respond well to a chelation regime is that the clinician does not recognize this and/or does not adjust the dose downward soon enough to prevent toxicity from over chelation. This is more likely in centres without long term or regular experience in monitoring and prescribing iron chelation. Regular monitoring for SF trends (1-3 times monthly) and for the known toxicities of each chelator are minimum requirements. The general principle of downward dose adjustment with rapidly falling body iron loads is clear, but the specifics as regards how much and when are less clear with some of the newer chelator regimes, and these are discussed in the respective sections on individual chelators. In general, the risk of over chelation with DFO increases when the SF is low relative to the dose. This has not been analysed systematically with other chelation regimes. With DFX, low levels of SF can be achieved even in patients not receiving transfusion, provided the doses are low (5-10 mg/kg), as SF values fall below 500µg/L (Taher 2013). Cases of toxicity from over chelation have been observed when SF are higher than this, if the rate of decrease is rapid. DFX dose adjustment consideration should be made at the first sign of increasing serum creatinine values (as a late event associated with falling SF or LIV values). With DFP it is not clear whether to or how to adjust dosing at low levels of SF or with rapid decrements in SF.

Unwanted effects with desferrioxamine

Unwanted effects with DFP and their management

• General tolerability and frequency of adverse effects
  In clinical trials, the rates of adverse reactions based on pooled data collected from 642 patients who participated in single arm or active-controlled clinical studies are available. The most common were chromaturia (red in colour due to urine in iron), nausea (13%), vomiting, abdominal pain (10%), elevations in alanine aminotransferase (8%), arthralgia (10%) and neutropenia (7 %). Other unwanted effects >1% were back pain (2%), arthropathy (1%), agranulocytosis (1.7%) change in appetite (5%). diarrhea (3%), dyspepsia (2%) and headache (3%) (FDA 2011).
• Relationship to dose and levels of iron overload
  Most studies where tolerability has been reported have used 75 mg/kg in 3 divided doses. The drug is licensed up to 100 mg/kg/day but insufficient numbers have been reported to know whether the incidence of the most serious complication, namely agranulocytosis, is increased at these higher doses or indeed decreased at lower doses. Because there are no formal studies examining dose in relation to unwanted effects in DFP, it is not clear which of the effects described below are due to over-chelation and which are independent of this mechanism. These are therefore described together. As a result of the various unwanted effects, 20-30% of patients are unable to sustain long-term treatment with DFP (Hoffbrand 1998).
• Neutropenia, agranulocytosis and thrombocytopenia
  The labeling for DFP includes a boxed warning stating that the drug can cause agranulocytosis (absolute neutrophil count or ANC <500/mm3), which can lead to serious infections and death as a consequence of infection. This agranulocytosis may be preceded by neutropenia. The frequency reported incidence is approximately 1.7% of patients. Each patient’s absolute neutrophil count should be measured before starting DFP therapy and weekly during treatment. The labeling states that DFP therapy should be interrupted and the patient’s neutrophil count closely monitored if an infection develops. The mechanism for DFP’s adverse effects on leukocytes is not known according to the labeling and is unpredictable in humans. Dose related effects on bone marrow hypoplasia and neutropenia has been seen in animal studies. Reported timing of onset of agranulocytosis is variable, from a few months to nine years. The condition may occur with thrombocytopenia, but isolated thrombocytopenia has also been occasionally reported, particularly in patients of Asian origin with probable hypersplenism.
  In a prospective trial where weekly neutrophil counts were undertaken and where DFP was discontinued when ANC was <1,500/mm3, agranulocytosis developed in 0.2/100 patient years and milder forms of neutropenia (ANC 500-1500/mm3) occurred in about 2.8/100 patient years (Cohen 2003, Cohen 2000). Neutropenia is more common in patients with intact spleens and commonly occurs in the first year of therapy. Recently, a continuation of DFP in patients with mild neutropenia (ANC 1000-1500/mm3) has been advocated by some investigators, with daily blood counts until resolution (El-Beshlawy 2013). However, it can be difficult in clinical practice to distinguish whether a low neutrophil count is part of a pattern of underlying neutropenia, or the first sign of agranulocytosis developing, which can be fatal. In 46 cases of agranulocytosis reported in Europe there were nine related deaths (Swedish Orphan 2006). Five cases were in patients who had been prescribed the drug for an unspecified ‘off label’ indications, and several were not receiving weekly blood count monitoring. Swedish Orphan subsequently issued the following recommendations on the use of DFP: ‘’ANC should be monitored every week or more frequently if there are signs of infection; concomitant treatment that could affect the white cell count should be avoided; if severe neutropenia or agranulocytosis develop, the drug should be stopped and not reintroduced, and the use of GM CSF should be considered in the case of agranulocytosis; off-label use of the drug should be avoided’’. Agranulocytosis and neutropenia has also been reported in patients taking combinations of DFP and DFO.
• Gastrointestinal symptoms
  Nausea, vomiting, gastric irritation and change in appetite (loss or gain) occurs in 3-24% of patients (Ceci 2002, Cohen 2000). A proportion of patients have to stop treatment for these reasons, which varies between studies. A liquid preparation is available which may be more palatable in children 1-10 years of age (ElAlfy 2010).
• Effects on liver
  In a summary of available clinical studies the FDA noted that 7.5% of 642 subjects developed increased ALT values. 0.62% of subjects discontinued the drug due to increased serum ALT levels, and 0.16% due to an increase in both ALT and AST (FDA, licensing information 2011). The frequency varied considerably between studies; about a quarter of patients show ALT fluctuations of twice the normal upper limit (Cohen 2000). By contrast, one prospective randomised study reported no significant end-of-study changes in liver enzymes with DFP or DFO (Pennell 2006b). An observational report of fibrosis after treatment of three or more years (Olivieri 1998) has not been supported by other reports (Wanless 2002, Hoffbrand, Tondury 1998). A relevant prospective randomised study investigating the progression to fibrosis using DFP for one year showed no difference as compared with DFO over the same period, and no difference in baseline and end-of-treatment liver function tests (Maggio 2002). A reduction of increased liver transaminase was observed in patients who were responding to DFP monotherapy, accompanied with a decreased LIC (Viprakasit 2013). Fluctuation of liver enzymes more than twice the upper limit of normal should prompt investigation of the cause and consideration for interrupting DFP therapy.
• Arthropathy
  The frequency of arthropathy varies greatly between studies, from as low as 4.5% at one year (Cohen 2000) to 15% after four years (Cohen 2003) in a predominantly European patient group, and as high as 33-40% in studies of patients from India (Sharma 2013, Choudhry 2004, Agarwal 1992). It is not yet clear whether these differences reflect environmental or genetic differences, or differences in iron overload between populations at the start of treatment. Symptoms range from mild non-progressive arthropathy, typically in the knees and controllable with non-steroidal anti-inflammatory drugs to (more rarely) severe erosive arthropathy that may progress even after treatment is stopped. Cases involving other joints, such as wrists, ankles and elbows and avascular necrosis of the hips have also been described. Radiologic evaluations revealed bony dysplasia, deformation and impaired growth of ulnar epiphyses in 13 out of 40 thalassaemia children who have received DFP for an average of 84 months (range 12-128 months) (Sharma 2013). It is recommended that treatment should be stopped where joint symptoms continue despite a reduction in DFP dose and are not controlled by non-steroidal anti-inflammatory drugs.
• Neurological effects
  DFP penetrates the blood brain barrier and has been evaluated for the treatment of conditions associated with brain iron deposition such as Freidrich’s ataxia (Tsou 2009). Neurological complications are very rare in thalassaemia treatment and have been typically associated with unintentional overdosing (>230 mg/kg/day). Reports of rare neurological effects have included stroke, cognitive effects, nystagmus, walking disorders, ataxia, dystonia and impaired psychomotor skills. These effects appear to improve on cessation of treatment. Cerebellar syndrome has been described with combination therapy.
• Effects on ears and eyes
  One study reported continued audiometric deterioration after switching from DFO to DFP (Chiodo 1997). A recent study found hearing impairment and audiometric abnormalities in 56% of children with TM receiving DFP or DFO, with no difference between the two chelation groups, with the main shared risk factor low ferritin (Chao 2013). It may therefore be advisable to monitor audiometric function in patients on regimes containing DFP as well as DFO. There have been isolated reports of loss of vision (central scotoma). A regular eye examination including retinal evaluation at least once yearly is therefore advisable.
• Other effects
  Zinc deficiency has also been observed in some patients, especially those with diabetes (Al-Refaie 1994). Zinc deficiency is difficult to measure in plasma samples, and needs to be taken during fasting, in the absence of chelator in the blood. Zinc deficiency has been linked to toxicity of DFP in animal studies and were abrogated by zinc supplementation (Maclean 2001). Some clinicians routinely add zinc supplementation with DFP monotherapy or combination therapy is given (not given at the same time as the DFP) (Porter 2013b).
• Frequency of adverse events compared with DFO
  Adverse effects have been reported in four randomised studies comparing DFP with DFO. One trial has reported data that allows comparison of the probability of an adverse event with DFP and DFO (Maggio 2002), establishing a statistically significant two-fold difference between DFP (34%) and DFO (15%), but no difference between temporary or permanent treatment withdrawal.
• Pregnancy
  DFP is teratogenic in animals and must never be given to patients attempting to conceive. Until more is known, potentially fertile sexually active women and men taking DFP must use contraception. DFP should not be used in pregnant women.
• Post marketing experience
  A number of additional adverse reactions have been reported with post marketing experience (see highlights of prescribing information in section 6.2 (FDA 2011)). Because these are reported voluntarily from a population of uncertain size, it is not always possible to reliably estimate their frequency or to establish a causal relationship to drug exposure.

Unwanted effects with DFX and their management

• General tolerability and frequency of adverse effects
  DFX has been given in the context of prospective trials to over 5900 patients with over 5 years of follow up in some prospective studies, so the relationship between DFX and unwanted effects are relatively well documented and defined. The frequency of adverse effects appears somewhat higher at doses of 25 to < 35 mg/k g/day (n = 136; 39.4%) than at 15 to < 25 mg/kg/day (n = 118; 31.1%) or < 15 mg/kg/day (n = 101; 29.4%).
• Relationship to dose and iron overload
  In general it is helpful to divide these effects into those which result from excess dosing with respect to the degree of iron overload and those which do not. Gastrointestinal, skin and renal effects can all be affected by dosing, although the exact relationship to body iron load has not been determined. It has however been shown in NTDT that with suitably low dosing (not exceeding 5-10 mg/kg) and low levels of ferritin or LIC, that low ferritin and LIC levels can be achieved without renal or other toxicities. Although a therapeutic index (as with DFO) has not been calculated, the general principle of reducing the dose, either as serum ferritin falls rapidly, or drops below 1000 µg/L for the first time would be adhered to. Although drug labelling suggests interrupting dosing when the serum ferritin reaches 500 µg/L, this leads to a ‘stop-start’ approach, which risks rebounding of NTBI and labile iron pools. Many clinicians therefore operate a dose reduction policy; giving very low doses (5 mg-10 mg/kg) to those patients who continue to be transfused. Some effects are notable by their absence, such as effects on growth, bone and arthropathy. The relationship of unwanted effects to iron load (LIC) has also been reported in 373 patients (Porter 2013a). Here, drug-related gastrointestinal AEs - mostly mild to moderate - were more frequently reported in patients with baseline LIC <7 versus those >/=7 mg/g dry wt, and were not confounded by diagnosis, dosing, ethnicity or a history of hepatitis B or C. Reported serum creatinine increases did not increase in low versus high-iron cohort patients. The smaller ESCALATOR trial however, found no clear trends in the type or frequency of drug-related AEs between the LIC <7 and ≥7 mg/g dry wt cohorts (Taher 2009). Tolerability of DFX tends to improve with long-term treatment (Cappellini 2011, Cappellini 2010). A practical guidance paper has been published which includes a summary of DFX -associated adverse events and a set of proposed management strategies (Vichinsky 2008).
• Gastrointestinal effects
  Gastrointestinal events are relatively frequent with DFX therapy but are typically mild to moderate and include diarrhoea, abdominal pain, nausea and vomiting, occurring in approximately 15-26% of patients (Vichinsky 2010). These symptoms rarely require dose adjustment or discontinuation, and decrease year on year over 5 years of follow up (Cappellini 2011). In the EPIC study, these symptoms were more common in patients with low baseline LIC values, as were abnormal LFTs (Porter 2013a). It is unclear to what extent the lactose component of the DFX formulation affects gastrointestinal tolerability in lactose-intolerant patients but this requires clarification, as lactose intolerance is common - particularly in South-East Asia. The role of co-administration of acidophilus or lactobacillus probiotic yoghurt to aid lactose has not been systematically studied. Hematemesis and melena due to gastric and/or duodenal ulceration has been reported in patients taking DFX (Yadav 2013, Bauters 2010). These patients should be investigated and managed appropriately, including Helicobacter pylori eradication therapy if required. Special attention should be taken in patients taking concomitant medications that can increase the possibility of gastric ulceration. Although the manufacturer does not recommend taking DFX with food or dividing the doses, some clinical experience has found that more flexible dosing schedules may be more suitable in patients who experience gastrointestinal disturbances. While the drug should still be taken at the same time each day, administration of DFX in the evening with or after food can potentially improve gastrointestinal tolerability and compliance. For example, a recent study found that palatability and gastrointestinal tolerability of DFX was improved when patients were allowed to take treatment with a soft food at breakfast, or with a beverage of choice (Goldberg 2013). Another option is to divide the DFX dose, which can help to reduce gastrointestinal events without adversely affecting iron excretion (Otto-Duessel 2007).
• Skin rashes
  Skin rashes occurred in 7-11% of patients, and were typically pruritic, maculopapular and generalised, but occasionally confined to palms and soles of the feet. Skin rash is more common in Asian population (up to 18%), often mild in severity and rarely developing into severe drug-hypersensitivity (Viprakasit 2011). Rash typically develops within two weeks of starting treatment. A minority of patients require permanent discontinuation of therapy, and mild rashes often resolve without dose modification, and became very rare after year 1 of treatment (Cappellini 2011). For moderate to severe rashes, treatment should be stopped and later restarted at a very low dose (<5 mg/kg), slowly increasing to therapeutic doses. Severe skin rash associated with angioedema is rare, and unlike the more common rash seen with DFX, may not respond to interruption and reintroduction at a lower dose. Here, DFX therapy may need to be halted completely as the angioedema may be evidence of an immune sensitisation response.
• Renal effects
  • - An increase in serum creatinine ≥30% on at least two consecutive readings was observed in 38% of patients receiving DFX, most frequently at doses of 20 mg/kg and 30 mg/kg (Cappellini 2006). These increases were sometimes transient and generally within the normal ranges, never exceeding two times the upper limit of normal (ULN), and were more frequent in the population of patients having the most dramatic decrease in LIC and serum ferritin. In a randomized study, dose reduction of 33-50% was planned if at least two consecutive increases in serum creatinine were >33% above baseline. As the creatinine spontaneously normalised in a number of cases, dose reductions were instituted in only 13%. In about 25% of those cases, the creatinine then returned to baseline, while in the rest it remained stable or fluctuated between baseline and the maximum increase observed prior to dose reduction. At 5 years of follow up, no evidence of progressive renal dysfunction had been reported where the above doses and modifications were used (Cappellini 2011). Other causes of increasing creatinine should also be considered in patients on DFX therapy, such as renal stones or concomitant use of NSAIDs. If a patient becomes acutely unwell for another reason, such as septicaemic shock or severe acute vaso-occlusive complications in SCD, it is probably wise to interrupt chelation therapy until the general condition stabilises.
  • - Proteinuria may be present in about a quarter of thalassaemia major patients, irrespective of the underlying chelation therapy, with average values about three times that of healthy controls (Economou 2010). Elevation of urine calcium and cystatin C are also seen in patients on DFX or DFP and DFO, whereas elevation of B2microglobulin was seen in patients on DFX only (Economou 2010). It is recommended that urine is monitored regularly for protein, and this can be conveniently performed at the time of visits for cross matching blood. Although proteinuria can fluctuate considerably, if there is a clear upward trend in the protein/creatinine ration above 1 mg/g, interruption or dose reduction should be considered. Current drug labelling recommends monthly urine testing for protein, which is helpful in establishing trends in proteinuria, as isolated estimates can be misleading.
  • - Case reports of renal tubular acidosis (Fanconi syndrome) with electrolyte imbalance, and metabolic acidosis due to tubular dysfunction have been rarely reported in adults and children taking DFX (Rheault 2011, Grange 2010). All cases recovered following withdrawal of DFX and appropriate electrolyte supplementation. Symptoms of renal tubular acidosis can be non-specific but may include polyuria, polydipsia and dehydration. Investigations may show proteinuria, hypokalemia, hypophosphatemia, hyperchloremic metabolic acidosis with excessive loss of substances in the urine (e.g. amino acids, glucose, phosphate and bicarbonate). Some patients, especially children, have intercurrent infections associated with Fanconi syndrome. Renal impairment due to DFX may also develop as part of a generalized delayed hypersensitivity reaction. It is recommended that kidney and proximal tubular function be periodically monitored in patients receiving DFX throughout their course of therapy.
• Hepatic effects
  Generally speaking liver enzymes improve in line with falling LIC. However, increases in liver transaminases are occasionally seen. In the EPIC study 0.6% of 1115 TM patients showed an increase of AST >10x the upper limit of normal (Cappellini 2010). These changes are commonly observed within 1 month of initiating DFX therapy, although this may occur later, particularly at high doses in patients with low iron burdens. Careful baseline assessment of alanine transaminase levels (ALT) should therefore be performed before starting treatment, and monitored every two weeks for the first month. Thereafter, checking ALT approximately monthly is recommended. Abnormal liver function tests are more frequent in children receiving DFX, and in such instances chelation should be stopped and ALT levels carefully monitored to ensure they return to normal. Reintroducing DFX using a slow escalation schedule has been reported in such cases, where the patient is started on DFX 125 mg, with the dose increased every 2–3 weeks provided the ALT levels remain stable with weekly monitoring. Improvements in the liver pathology of 219 patients with beta-thalassaemia treated with DFX for at least 3 years has been reported in a prospective trial (Deugnier 2011). By the end of the study, stability of Ishak fibrosis staging scores (change of -1, 0, or +1) or improvements (change of </=-2) were observed in 82.6% of patients. DFX treatment for 3 or more years reversed or stabilized liver fibrosis in 83% of patients with iron-overloaded beta-thalassaemia.
• Effects on hematology
  Although hematological adverse effects such as thrombocytopenia and agranulocytosis have been added to the product information since DFX was approved, only a few cases have been published as case reports thus far. In two patients, thrombocytopenia due to DFX developed as part of a generalized delayed hypersensitivity reaction, in combination with a rash, fever, eosinophilia, hepatic and renal impairment (Wei 2011, Gutiérrez Macías 2010).
• Arthropathy and growth failure
  No cases of arthropathy or growth failure have been associated with DFX administration. Comparing 296 patients who received DFX in a one-year prospective randomised study with 290 patients receiving DFO, deafness, neurosensory deafness or hypoacusis were reported as adverse events in eight patients on DFX, and seven on DFO.
• Eyes and ears
  These are very rare and their significance is uncertain, however, current labelling recommends yearly auditory and eye assessments (Novartis 2013). Early lens opacity was reported in the DFX core registration trials, but the incidence (0.3%) did not significantly differ from the control group of DFO treated patients (Cappellini 2006). The electroretinographic effects previously seen with DFO have not been described, and the frequency at which electroretinography assessment is indicated has not been formally assessed. Possible audiometric effects were identified in early studies but this has not been reported systematically. One investigator has reported lens opacities in 3 out of 12 patients (Bloomfield 1978), which would approximate to 80-times the incidence observed in the large-scale trials, the reason for which is presently unclear (Ford 2008).
• Pregnancy and DFX
  DFX has been shown to have teratogenic effects in animal studies. However, there has been a report of a healthy male baby delivered to a 38-year-old mother with thalassemia major who has unintentionally conceived during DFX therapy (Anastasi 2011). However, at present it is recommended that thalassaemia patients who plan to conceive should avoid the use of iron chelation for at least 3 months before.
• Post marketing experience
  A number of additional adverse reactions have been reported with post marketing experience (see highlights of prescribing information in section 6.2 of (FDA 2011)). Because these are reported voluntarily from a population of uncertain size, it is not always possible to reliably estimate their frequency or to establish a causal relationship to drug exposure.

Practical issues with subcutaneous infusion

Because regular use of DFO is critical to a good outcome, every effort should be made with each individual to help him or her to find the most convenient way to infuse the drug.

• Strength of infusion
  The manufacturers of DFO recommend that each 500 mg vial of the drug is dissolved in at least 5 ml of water, giving a 10% solution. Concentrations in excess of this may increase the risk of local reactions at the site of infusion.
• Site of infusion
  Care must be taken to avoid inserting needles near important vessels, nerves or organs. The abdomen is generally the best place. However, because of local reactions such as erythema, swelling and induration, it is often necessary to ‘rotate’ the sites used for injection (see Figure 3). Some patients find that the skin over the deltoid or the lateral aspect of the thigh provides useful additional, alternative sites. The best needle to use will depend on the individual. Many patients are happy with butterfly needles of 25 gauge or smaller, which are inserted at an angle of about 45 degrees to the skin surface. The needle tip should move freely when the needle is waggled. Other patients prefer needles that are inserted vertically through the skin and are fixed with an adhesive tape attached to the needle (see Figure 4). Patient preference is highly variable and clinicians should explore the best type of needle for each patient, to help maximize compliance.
• Types of infuser
  There are many types of infuser now available. Newer devices, including balloon pumps, are smaller, lighter, and quieter than their predecessors. For patients who find dissolving, mixing and drawing up DFO a problem, pre-filled syringes or balloons may be useful. Some pumps are designed to monitor compliance.
• Local reactions
  Persistent local reactions may be reduced by varying injection sites, lowering the strength of infusion, or in severe cases, by adding 5-10 mg of hydrocortisone to the infusion mixture. Application of topical low potency corticosteroid cream after injection can reduce local reactions.

Figure 3.

Rotation of infusion sites.

Figure 4.

Insertion of needles for desferrioxamine infusions.

Practical details for intravenous infusions

10 % solutions of DFO given to peripheral veins will damage and sclerose the vein. If infused (as an emergency) into a peripheral vein, the solution must be diluted – for example in 200-500 mls of saline.

• Management of in-dwelling intravenous lines
  Infection and thrombosis of catheters may occur. Careful aseptic procedures must be followed in order to prevent possible infection by Staphylococcus epidermidis and aureus, which when established are difficult to eradicate, and often removal of the infusion system becomes necessary. The risk of thrombosis and infection is likely to be greater in centres that do not have regular experience in the use of long-term in-dwelling lines (Piga 2006). Use of prophylactic anticoagulation is advised, as line-thrombosis is relatively common in thalassaemia major (Davis 2000). As development of thrombosis can occur at the tip of the catheter, it is advisable if possible to avoid placing the tip in the right atrium.
• Intravenous DFO with blood transfusion
  This has been used as a supplement to conventional therapy (e.g. 1g over 4 hours piggy-backed into the infusion line), but its contribution to iron balance is very limited and not recommended as a standard procedure. Special attention must be given to avoiding accidental boluses due to DFO collecting in the dead space of the infusion line. Co-administration of DFO and blood can lead to errors in interpreting side effects, such as acute fever, rashes, anaphylaxis and hypotension during blood transfusion. DFO should never be added directly into the blood unit.
• Use of DFO by subcutaneous bolus
  If an infusion pump is not available or if 10-hour infusions are not tolerated, bolus subcutaneous treatment may be considered if the patient is not at high risk of heart disease. A randomised study has shown that serum ferritin and liver iron can be controlled equally effectively by giving an equivalent total dose (45 mg/kg × 5 per week) either as two subcutaneous ‘boluses’ or as a nightly 10-hour subcutaneous infusion (Yarali 2006). However, this technique may be impractical in the clinic – particularly in paediatric patients, due to the painful nature of bolus infusions.