Correspondence Sunil Bhandari, Department of Renal Medicine, Hull and East Yorkshire Hospitals NHS Trust, Anlaby Road, Kingston upon Hull, East Yorkshire HU32JZ, UK. Tel: (+44) 1482 674566; Fax: (+44) 1482 674998; e-mail: email@example.com
Beta-thalassaemia is characterised by progressive anaemia necessitating regular blood transfusions to sustain life. With the advent of effective chelating agents that can reduce the iron burden and extend patients' survival, renal disease, as in other ageing populations, has become more prevalent. In recent years, chronic kidney disease (CKD) has become overwhelming; indeed, approximately 8% or 6 million people of the UK population has evidence of CKD. Several factors, which occur in patients with thalassaemia, account for the relative explosion of renal disease in the general population including increasing age, diabetes, hypertension and the advent of novel measures of renal function facilitating early detection of kidney disease. In addition, some patients with thalassaemia develop renal tubular dysfunction related to the disease itself, the effects of iron overload and the effects of chelator therapy, while other patients have an increased creatinine clearance leading to hyperfiltration. More recently, there is a noticeable increasing prevalence of impaired renal function and proteinuria because of several putative factors including chelators. We review current data on the potential mechanisms leading to renal abnormalities seen in patients with thalassaemia, the potential effects of iron loading within the kidney and the potential renal effects of chelator therapy. This article gives a speculative account of possible mechanisms and theories to consider providing pause for thought and direct future research in this area.
Thalassaemia encompasses a spectrum of recessively autosomal hereditary anaemias characterised by reduced or absent production of one or more globin chains . Beta-thalassaemia results from the impaired production of beta-globin chains, leading to a relative excess of alpha-globin chains. These excess alpha-globin chains are unstable and cannot form soluble tetramers on their own, and therefore precipitate within the erythroid precursors in the bone marrow leading to their premature death (ineffective erythropoiesis). The degree of alpha-globin chain excess determines the severity of subsequent clinical manifestations, which are profound in patients homozygous for impaired beta-globin synthesis and much less pronounced in heterozygotes .
At a clinical level, three main forms have been described: the severe transfusion-dependent thalassaemia major, the milder not transfusion-dependent thalassaemia intermedia (a result of different complex genotypic combinations) and the asymptomatic carrier.
Homozygous beta-thalassaemia is usually characterised by progressive anaemia from birth necessitating regular blood transfusions to sustain life apart from those with thalassaemia intermedia . As a consequence of chronic transfusion, iron accumulates in various organs and tissues, resulting in progressive dysfunction of the liver, endocrine glands and heart . Historically, renal disease has not been a major issue in patients with thalassaemia because survival was limited by severe cardiac iron loading from chronic transfusion therapy leading to premature early death [4-6], and simply, patients did not live long enough to develop conditions linked to kidney dysfunction (e.g., diabetes). With the advent of effective chelating agents that can reduce the iron burden and its consequences, and extending patients' survival, renal disease, as in other ageing populations, potentially may become a more common occurrence.
In recent years, chronic kidney disease (CKD) in the general population has become overwhelming; indeed, approximately 8% of the UK population which equates to approximately 6 million people, and 13% of the US population has evidence of chronic kidney disease [7-9]. This presents a potentially huge economic burden for society to manage. Several major factors account for the relative explosion of renal disease in the general population. These include the increasing age of the population, including patients with thalassaemia, diabetes which has become a global epidemic and is projected to increase further by 46%, increasing hypertension and obesity. All these factors occur, except perhaps hypertension, to a greater or lesser extent in the ageing thalassaemia population. Also, clinicians, with the advent of estimated glomerular filtration rate (eGFR) and other novel measures of renal function (cystatin C), are becoming more aware of kidney disease as a major medical problem, which needs proactive measures to prevent future problems [10, 11]. In thalassaemia major, there are no prospective or longitudinal studies examining renal disease. Historical data have shown that thalassaemia has represented a small proportion of dialysis patients. Indeed, significant renal involvement is not a frequent complication in children and young adults suffering from thalassaemia. Clearly, however, these estimates may change with the future ageing thalassaemic population and use of medications, which may potentially affect the kidney.
This review therefore summarises the historical and most recent current data on the potential mechanisms leading to renal abnormalities seen in patients with thalassaemia, the potential effects of iron loading within the kidney and the potential renal effects of chelator therapy. Owing to the lack of quality clinical data, theories and speculation to possible mechanisms and clinical sequelae are detailed where appropriate to allow the reader/clinician to consider the significance and relevance of renal disease in patients with thalassaemia in clinical practice and direct future research in this field. The most recent message from the World Kidney Day stressed that an ‘increased attention to the kidneys will indeed improve long-term health outcomes by reducing both kidney and cardiovascular disease' .
Measurement of renal function
Serum creatinine is recognised as an unreliable measure of early renal dysfunction. Other markers such as eGFR based on the Modified Diet in Renal Disease (MDRD) four variable equation in adults and Schwartz equation in children are more reliable while cystatin C remains controversial . These, however, remain to be validated in patients with thalassaemia. Papassotiriou et al. and others have demonstrated that cystatin C is a poor biomarker of renal function and although it is raised in up to 60% of patients on deferasirox (one of the three available iron chelators see below), it appeared to correlate inversely with serum ferritin and was more reflective of changes in cardiac function . This fits with data suggesting a stronger association of cystatin C with cardiovascular outcomes and potentially overall mortality [15, 16].
Mechanisms leading to renal abnormalities in thalassaemia
Studies in various thalassaemia populations more than a decade ago, one Italian (n = 7731, thalassaemia carriers) and the other Canadian (n = 216, 188 beta-thalassaemia minor and 27 haemoglobin H or E beta-thalassaemia), have shown 0.5% of patients developed renal tubular dysfunction and 3.1% (n = 240) of patients progressed to dialysis therapy in the former cohort and 8% had a reduced creatinine clearance and 21% had an increased creatinine clearance . A more recent cross-sectional study in all thalassaemic groups including thalassaemia major has confirmed these findings with an increased creatinine clearance in 20.8%, impaired renal function in 7.8% and albuminuria in up to 59% of cases . However, this data do not tease out the degree of influence of chelators in relation to dialysis requirements and renal dysfunction and also mainly applies to patients with thalassaemia minor.
Renal tubular function
Renal tubular functional abnormalities have been described with increasing frequency in patients with beta-thalassaemia major . However, the extent of intrinsic renal injury in patients with thalassaemia is poorly defined. Damage to tubular cells may manifest in several ways from simple protein leak in the urine on dipstick to more severe damage leading to proximal tubular dysfunction and potentially acute tubular necrosis . A variety of renal tubular abnormalities have been described in patients with thalassaemia major: they include hypercalciuria (22%), hyperuricosuria (54%) with renal uric acid wasting, renal glycosuria and tubular proteinuria [20-23]. Haematuria is relatively common and may relate to an increased incidence of renal stone disease directly causing haematuria from trauma to the renal tract . This increased incidence of renal stones is related to the increased uric acid in the urine as a consequence of high erythrocyte turnover and the increased hypercalciuria from vitamin D and calcium replacement for deficiency and hypoparathyroidism [23, 24]. Therefore, this may lead to an increased likelihood of acute and chronic obstruction, and recurrent infections and thus progressive renal injury. Therefore, an acute deterioration in renal function should alert the clinician to the possibilities of renal stones especially when clinical symptoms or signs (dipstick positive blood) of renal colic are present. The incidence of renal stones is not common; in the general population, 1% will suffer a clinical episode related to renal stones, while in some cases, up to 8% of patients with thalassaemia major, at least in one series, were found to have a stone . If stones or haematuria are present, then patients should undergo cystoscopy and further evaluation to characterise the nature of the stone and possible metabolic risk factors increasing the risk of stone formation such as excessive vitamin D replacement.
Many studies in thalassaemia, including major, have identified an increase in biomarkers of proximal tubular damage, leading to increased urinary excretion of N-acetyl-d-glucosaminidase (NAG) (found in lysosomes of tubular cells and putatively early marker reflecting proximal tubular cell necrosis), beta-2 microglobulin, malondialdehyde (MDA), amino acids and low molecular weight proteins (Table 1) [22, 25-27]. Therefore, one can surmise from these studies that there is a direct association between thalassaemia and tubular dysfunction, which may be causal, but as yet unproven. A possible mechanism to account for these findings is perhaps that thalassaemia itself may lead to proximal tubular dysfunction either through chronic hypoxia from persistent anaemia or through haemosiderosis/iron deposition; however, studies examining this phenomenon have been performed with healthy children as controls rather than non-chelated children with thalassaemia [19, 23, 25].
Table 1. Prevalence of markers of renal glomerular and tubular-related effects in patients with thalassaemia
Hypercalciuria > 0.25
Excretion of beta-2 microglobulin
Proximal tubular defects
In patients with thalassaemia major, hyperfiltration has been noted in several studies [19, 20, 24, 25, 27, 28]. This may mimic the early changes seen in diabetic nephropathy. Renal hyperfiltration, if untreated, leads to progressive tubule-interstitial injury and glomerulosclerosis from disruption of the renal architecture, through effects on mesangial cell function and increased proteinuria. This process is driven by several factors including haemolysis which in turn affects nitric oxide–dependent vasodilatation, the effects of anaemia with reduced peripheral and renal vascular resistance and hence increased renal plasma flow, and finally iron deposition in the renal parenchyma. However, thalassaemia major is a disorder characterised by ineffective erythropoiesis, thus, haemolysis in the periphery is generally not a prominent feature of this disorder, particularly in those who are receiving regular transfusions that also suppress erythropoiesis. Thus, the impact of nitric oxide scavenging on glomerular filtration may not be as prominent leaving the other factors to perhaps lead to the changes. Therefore, these high eGFR values may have an adverse prognosis and apply to all clinical disorders including thalassaemia .
Acute kidney injury (AKI) has been reported in thalassaemia . The probable mechanism in many historical cases is probably prerenal from sepsis or complications of heart failure (the cardiorenal syndrome) and liver failure affecting renal perfusion. Prerenal causes from drug side effects such as diarrhoea and vomiting, idiosyncratic or toxic effects may occur. However, from the current literature, there are few reports of the latter outcomes.
De novo renal disease may also occur and exacerbate glomerular dysfunction. Hepatitis B and C, which are relatively common in older thalassaemia patients, should be remembered as potential causes of renal disease (including membranous and mesangiocapillary glomerulonephritis). IgA nephropathy has previously been described in a patient with thalassaemia . Previous animal studies have shown that bleomycin-detectable iron (iron capable of catalysing free-radical reactions) was markedly increased in glomeruli from nephrotic rats.
Shah et al.  have also demonstrated that in patients with overt diabetes or microalbuminuria, there is a marked increase in urinary catalytic iron. Interestingly, in preliminary studies, it has been shown that treatment with deferiprone leads to a marked reduction in proteinuria over a 9-month period in patients with diabetic nephropathy. Recent experimental data in the prototypical mouse model of human type II diabetes Leprdb Leprodb mouse have shown that 125 mg/kg deferiprone resulted in a significant reduction in 24-h urinary proteinuria (943 to 529 μg/mg; P = 0.032) and a modest improvement in serum creatinine . These findings suggest iron itself, in the catalytic form rather than bound iron as demonstrated by Agarwal,  is potentially toxic to the renal parenchyma.
The potential effects of iron loading on the kidney
Post-mortem studies demonstrate haemosiderin deposition in visceral and parietal glomerular epithelial cells in both proximal and distal convoluted tubules leading to the possibility of tubular dysfunction . Also from studies in iron overloaded rats, it is possible that high levels of iron are sufficient to cause cellular damage via stimulation of reactive oxidative elements . How strong this evidence for iron-induced tubular dysfunction stems from data from Aldudak and colleagues who have demonstrated higher levels of MDA consistent with lipid peroxidation induced renal tubular injury  This might result from increased tissue iron but data on this are sparse and chelator therapy appears to reduce levels of MDA production . There may, however, be a direct effect of iron on proximal tubular function. Although there is some correlation of ferritin with urinary excretion of NAG and beta-2 microglobulin , correlation of tubular changes with ferritin levels is less clear . Indeed, evidence shows a 31% incidence of generalised aminoaciduria, with similar results among chelated and non-chelated groups . A recent case–control study of 140 patients with thalassaemia major aged between 7 and 16 yr who had been exposed to regular blood transfusions (1–2 per month from an early age) demonstrated a significant increase in urinary NAG in comparison with controls, with 58% of cases having a high NAG . This increased with age, and there was a positive correlation of serum ferritin with NAG. However, there was no correlation of serum ferritin with eGFR. There was a trend to a falling eGFR with age: only five of 93 cases had an eGFR <100ml/min/1.73m2. It would appear that, this study confirms those previously carried out by Mohkam et al.  and suggests that number and duration of exposure to blood transfusions and hence iron load are important factors in renal tubular toxicity. Despite this recent study, whether this effect of iron leads to renal progression has not been studied, as most data emanate from children over a short time span  and focus mainly on tubular function. Future long-term studies are needed to clarify the effects of iron on tubular function.
In order for iron to function in biological processes, it has to cycle reversibly between its ferrous and ferric oxidation states. This property makes it potentially toxic, because free or catalytic iron leads to free-radical generation, which can subsequently damage cells and affects cell function. In thalassaemia, this potential transitional pool of free iron (extracellular and intracellular) may play a major role in lipid peroxidation from catalysing the formation of reactive oxygen species, leading to injury of the lipid bilayers or organelle membranes [37, 38] The net effect of this renal iron manifests as proximal tubular dysfunction.
Labile iron may also lead to AKI [39, 40], which has been defined as abrupt reduction in kidney function of the order of an absolute increase in serum creatinine of at least 26.4 micromol/L, or a percentage increase in serum creatinine of at least 50% from baseline. In patients with non-thalassaemia, there is evidence of iron-related toxicity to the kidney [32, 41, 42]. Several authors have shown that in myoglobinuric AKI, there is a marked increase in labile iron content [39, 40]. They have demonstrated that iron chelation had a protective effect on renal function and was also associated with a marked reduction in histological evidence of renal damage [39, 40]. Exposure of kidney cells to cisplatin leads to increased labile iron, while chelation attenuated cisplatin-induced cytotoxicity [39, 40]. In addition, gentamicin, at least in vitro, leads to generation of hydroxyl radicals and hence hydrogen peroxide with release of iron from cortical mitochondria. More recently, Rajapurkur et al.  have reported that kidney donors undergoing either an intravenous pyelogram or a renal arteriogram have a marked increase in urinary catalytic iron accompanied by evidence of tubular injury. Animal models of contrast-induced AKI with addition of an iron chelator has been shown to be protective . These findings, however, are limited to small non-controlled studies.
A recent study measuring iron in the kidney of patients with sickle cell disease has shown minimal deposition in comparison with controls . Indeed, in the same study, the high R2* (reciprocal of T2*) iron signal was mainly as a result of haemolysis causing the high renal iron content [45, 46]. Therefore, the process is perhaps more complex involving a number of inter-related pathways (Fig. 1). But clearly in the kidney, it may not be a dose-related effect. Indeed, in CKD, Ferrari et al.  have shown that iron loading with exogenous intravenous iron results in a transient deposition in the liver, indicating that renal failure itself does not significantly influence the effect. This theory seems plausible as current data suggest that iron loading per se is not the culprit. Future studies are needed but preliminary clinical studies from Argawal et al.  have shown that intravenous iron increases oxidative stress markers because of the release of labile-free iron.
The potential effects of thalassaemia-related chronic hypoxia and anaemia
In considering the potential mechanisms of renal injury, anaemia and associated potential chronic hypoxia could be important (Fig. 1). Anaemia can lead to activation of the oxidative stress cascade once again with the end result of lipid peroxidation and cell damage and eventual functional change of the tubules . The anaemia may also lead to changes in the morphology of cells in terms of size and vascular supply . Therefore, clinical caution is warranted with recommended monitoring for renal tubular and glomerular function with urinary dipstick testing and measurement of serum creatinine and eGFR as a minimum.
Potential effects of chelation therapy
All drugs are potentially toxic. Indeed, certain herbal remedies are linked to kidney failure. For prescribed medications, there are numerous potential effects at various levels in the kidney, from the inherent toxicity of the molecule, the potential adverse effects leading to renal dysfunction, effects on renal blood flow, effects on signalling pathways, idiosyncratic effects and direct cell damage (toxicity) of a drug on the kidneys. These all warrant consideration as potential hazards of therapy.
Chelation therapy is an established treatment for heavy-metal poisoning. Heavy metals, which cannot be metabolised, physiologically excreted, persist in the body and exert their toxic effects by combining with one or more reactive groups (ligands), essential for normal physiological functions. Many chelating agents exist and have been used in the treatment of many metal poisoning disorders and aluminium toxicity in renal failure. Chelators are been used for the treatment of chronic iron overload in patients with thalassaemia major.
Three iron chelators are currently available: Deferoxamine, in use since the 1960s mainly given subcutaneously for 10–12 h/d, for 5–7 d/wk, has an excellent safety and efficacy profile ; Deferiprone (Ferriprox), introduced in clinical use in India in 1995 and in Europe in 1999 and most recently North American in 2011 and given orally three times per day, has shown particularly beneficial effects in reducing cardiac iron and improving cardiac function and outcomes [5, 51, 52]; and Deferasirox (Exjade), introduced in 2005 in North America, and in 2006 in Europe, given as oral suspension once a day, has shown excellent results in clinical trials in its ability to reduce liver iron concentrations and serum ferritin .
All three chelators can potentially lead to consequences on renal function from prerenal effects, as a result of volume depletion from diarrhoea and vomiting-related adverse events, to haemodynamic and specific nephrotoxicity (Table 2). There are little data verifying this potential outcome but it should be considered by clinicians if an episode of AKI occurs. It has been suggested that the rapid iron chelation, as evidenced by a rapid drop in ferritin values, correlates with renal dysfunction [54, 55]. In addition, the presence of other co-morbidities such as type 2 diabetes and essential hypertension may lead to a higher susceptibility to the development of renal side effects. However, unravelling the mechanisms of renal injury remains challenging as up to approximately 25% of patients use combination therapy for optimal iron removal.
Table 2. Potential renal toxicity related to iron chelators
Increase in serumcreatinine Transient or persistent
Yes – occasional
Acute kidney injury
– fatal or non-fatal
None to date
Yes – occasional
Yes – rare
Proteinuria > 0.5 g/d
Yes – uncommon
Renal tubular acidosis
Hyperchloraemic metabolic acidosis
Yes – uncommon
Deferoxamine use in high doses is now infrequent and renal dysfunction less common. AKI and acute changes in renal function have been reported in up to 40% of patients given the drug, historically when higher doses and overdoses occurred [24, 56-63]. In previous randomised studies in thalassaemia, 14% of the cohort given deferoxamine demonstrated increases in serum creatinine. This might suggest that the effect may in part be due to iron depletion leading to effects on renal function via impaired mitochondrial function [53, 64]. The increased beta-2 microglobulin and non-oliguric AKI [24, 57] might suggest tubular damage from deferoxamine but interestingly deferoxamine affects renal blood flow and GFR, leading to increases in serum creatinine especially at higher doses or normal doses in higher risk patients (diabetics, patients with hypertension, patients with proteinuria, elderly patients and patients with underlying renal dysfunction)  with recovery on discontinuation. Renal histology in a single case of AKI confirmed acute tubular injury and significant damage to their mitochondria . Jalali et al.'s  recent study also indicates that exposure to deferoxamine poses a risk of renal toxicity, in part related to repeated iron transfusions or indeed iron depletion. The relatively high incidence of Yersinia infection with deferoxamine therapy poses a potential cause if sepsis follows [65-67]. Although not reported in the literature, the potential of this infection, if not promptly treated to cause AKI, should be considered. This may be particularly important in at risk groups with renal/cardiovascular risk factors . Most recently, Aydinok et al.  in an observational study of children under 5 yrs exposed to deferoxamine for a median 2.3 yrs demonstrated that 5.1% of patients developed non-progressive increases in serum creatinine and 5% had a >50% increase in serum creatinine values and 2.5% having a value greater than the upper limit of normal. Interestingly, review of the graphs provided show an increase in median serum creatinine from approximately 26.4 to 34.8 micromol/L over the 2-yr follow-up. The short follow-up raises the question of the definition of non-progressors taking into consideration age-related changes in renal function.
Oral deferiprone, from the current available literature and drug surveillance studies, seems safe in relation to effects on the kidney. Data are limited in high-risk groups, and there are no prospective studies to refer to. We have not experienced significant transient and/or persistent renal adverse effects with using deferiprone in a small cohort of 70 patients with thalassaemia major in Oman . However, deferiprone may cause, in about 1% of the treated patients, agranulocytosis requiring drug cessation . This latter side effect could potentially cause AKI if severe sepsis, which has been reported in some cases of deferiprone-induced agranulocytosis, leads to hemodynamic decomposition and acute tubular necrosis. This is rare but should be considered by clinicians as sepsis in general accounts for a large percentage of AKI presentations to hospital. El Alfy et al.  from a study of 100 children under 10 yrs (mean 5.1 ± 2.4 yrs) with transfusion-dependent anaemia (91 patients with thalassaemia major) and given liquid deferiprone demonstrated a modest non-significant and non-progressive increase in serum creatinine from a baseline mean of 29.2 ± 12 to 37.1 ± 10 micromol/L over the 6-month period. Two patients had baseline serum creatinine values of 71 and 62 micromol/L, which resolved by the end of the study, while two others had a single episode of elevation of serum creatinine above the upper limit of normal, but both also resolved by the end of the study without a need for interruption of therapy or dose adjustment . The reason for these fluctuations is unclear but has not been seen in other currently available data.
Oral deferasirox has become a routine therapy for the treatment of iron overload in patients with transfusion-dependent anaemia. Several renal side effects have been reported with increasing frequency [24, 58, 59, 62, 73-77]. Cases of reversible mild or even life-threatening Fanconi syndrome, in both children and adult patients, have mainly been associated with patients receiving prolonged (>6 months) courses of deferasirox. Recovery is usual upon complete withdrawal of deferasirox but recurrence of Fanconi syndrome even after initiation of lower doses of deferasirox may occur . The rapid improvement in biochemical parameters after stopping the drug and the lack of other potential aetiologies strongly points to deferasirox as the cause of Fanconi syndrome. Perhaps, the two iron (III) complexes of deferasirox, which are non-charged, less readily penetrate membranes in comparison with other chelators. A comparative study of the permeability of Caco-2 cell monolayers to the iron complexes of the available 3 chelators has indicated that only deferiprone was found to be able to cross the monolayer . Thus, although deferasirox is able to readily enter cells because of its lipophilicity, it forms a highly charged complex with iron, which one would predict will not readily efflux from cells. This triple negative charge of the deferasirox-iron complex may therefore in part explain the nephrotoxic potential of deferasirox.
In a phase 3 multi-centre study of 586 paediatric and adult patients (mean age, 17 yr; range, 2–53 yrs) taking deferasirox (n = 296) compared to deferoxamine (n = 290) over a 1-yr period, 38% receiving deferasirox developed an undefined dose-dependent transient and usually non-progressive increases in serum creatinine, in particular in those with a ‘dramatic decrease in liver iron content and serum ferritin’ . In approximately 25% of these patients, the creatinine returned to baseline without changes in therapy, while in the remainder, the creatinine was stable. Dose reduction, temporary medication interruption or discontinuation were successful to avoid progressive renal dysfunction. Proteinuria was infrequent (0.6%) [53, 64]. In a study of 59 paediatric patients receiving deferasirox, creatinine increases were described as mild and remained within the normal laboratory range . Mild increases in serum creatinine were more common in patients with the greatest reductions in serum ferritin . Again from the observational data from Aydinok et al. in children under 5 yrs exposed to deferasirox for a median 2.3 yrs, 18% of patients developed non-progressive increases in serum creatinine, 14% had a >50% increase in serum creatinine values and 4% having a value greater than the upper limit of normal . Review of the graphs provided show an increase in median serum creatinine from 19.4 to 35.2 micromol/L over the 2-yr follow-up. Interestingly, no dose adjustment of a median dose of 26.2–29.4 mg/kg/d was required . This stable renal function is consistent with recent data with higher doses of deferasirox [79, 80].
The EPIC cardiac study has since progressed to 3–5 yrs and a sub-group cohort of 114 patients suggest safety of use over 5 yrs, with no deaths in patients who, at baseline, had normal renal function and no risk factors such as diabetes, proteinuria or hypertension. Indeed, there were few episodes (2.3%) of significant rises in serum creatinine in subsequent years . Interstitial nephritis has been reported in patients with myelodysplastic syndrome on deferasirox therapy, but similar to commonly prescribed antibiotics, this effect is unpredictable in patients and resolves with medication discontinuation [62, 81-83].
Mechanisms of renal injury with chelation
The mechanism of renal injury associated with chelation is unknown. AKI with chelators  is unexplained but recognised. In preclinical studies in primates, deferasirox caused a 95% reduction in liver iron content with a 40% reduction in kidney iron content . A similar phenomenon may occur with deferoxamine. The chelation of proximal tubule mitochondrial iron with resulting ATP depletion has been postulated to be associated with proximal tubular dysfunction. In the case of deferasirox-associated iron chelation therapy, the observed increase in creatinine thought to be due to reduction in glomerular filtration rate and renal blood flow in some patients . Recent prospective data (n = 10 patients with thalassaemia major) have indicated a potential vascular mechanism for the transient 22% rises in serum creatinine and 17% reduction in creatinine clearance, with a 9–21% reduction in renal blood flow after 8 wks of therapy . These parameters all normalised during a 2-wk washout period of drug discontinuation. Therefore, this apparent nephrotoxicity is in part haemodynamic, reversible and may account for some of the cases of AKI and would appear safe in low-risk patients with no underlying renal disease. The mechanism for this change in renal blood flow might in part relate to chelator-induced iron depletion, affecting the afferent arteriole via an upregulation of prostaglandin production and arteriole vasoconstriction . However, proteinuria can occur in up to 18% of patients on deferasirox or deferoxamine but not deferiprone currently. One might envisage that reduction in renal blood flow would reduce this effect initially. Therefore, the mechanism would appear to be more complex and perhaps be beneficial for the hyperfiltration. Longer term data continue to emerge but data from long-term exposure on the renal effects are lacking. Importantly, recent data from post hoc analysis of the RENAAL study of diabetic patients with CKD suggested that the initial fall in eGFR with renin angiotensin aldosterone blockade was haemodynamic affecting intraglomerular pressures rather than structural, but interestingly longer term suggested benefit . It would be intriguing to consider a similar effect with deferasirox in patients with thalassaemia where the initial falls in eGFR may be beneficial in the longer term on renal function. However this needs to be weighed alongside the tubular toxic effects displayed with deferasirox.
The long-term consequences of proximal tubular dysfunction in patients treated with chelators are unknown and deserve further study. Observational studies, both in vitro and in vivo, of deferasirox suggest that toxicity appears to be dose related and affects biomarkers. From in vitro work, one can conclude that the doses used to produce a Cmax equivalent to 30–40 mg/kg/d are sufficient to be potentially toxic to cells . Teasing out this mechanism is challenging. The free iron liberated within cells on chelation (trapping of iron-bound deferasirox within tubular cells within the kidney) and iron excess may have both direct and indirect effects. In CKD, parenteral iron sucrose therapy has been shown to lead to increased proteinuria possibly through inflammatory effects, oxidative activation and indeed, at least in animal studies, deposition in mesangial cells and podocytes is seen [42, 89]. The iron may directly affect the mitochondrial membrane pore opening with a resultant reduced respiratory capacity, while indirectly via iron-catalysed oxidant damage of mitochondrial DNA, which is more vulnerable, leading to disruption of the respiratory chain complexes I to IV and thus leading to failure of mitochondrial respiration and cell damage. Also, a series of metabolic (reduction in fatty acid oxidation and a subsequent failure to remove potential toxic lipids) and cellular changes may be important in causing dysfunction. Reduced ATP synthesis affects the mitochondrial pore while affects on PI3K/AKT lead to apoptosis with resultant cell and mitochondrial death [68, 90]. Finally, mutations to mitochondrial DNA may result from the effects of the milieu surrounding the cells . The end clinical result is cardiac failure or tubular damage. There are differences in mitochondrial function along the nephron segments, which may differently affect damage – proximal tubules tend to be more oxidised than distal tubules and hence more susceptible.
Thalassaemia and renal disease may occur more commonly in association as the population ages. Renal disease is a surrogate and independent marker for an increased risk of cardiovascular events (strokes, heart failure and myocardial infarction), which increases the risk of mortality . Indeed, a person with <45 ml/min kidney function, based on eGFR, compared to a person with >75 ml/min kidney function has an almost fivefold increased risk of having a myocardial infarction or stroke or death [15, 93]. The presence of proteinuria/albuminuria leads to a fourfold increased risk of having a cardiovascular event or developing future kidney dysfunction. Therefore, these two factors, the presence of kidney dysfunction and the presence of proteinuria are synergistic, causing a multiple increased risk on life to patients with thalassaemia. Interventions such as angiotensin-converting enzyme inhibitors can reduce the risk by over 44% .
Renal insufficiency in thalassaemia
Although renal disease has not been a large problem in thalassaemia, it is perhaps inevitable that more patients with thalassaemia may develop multiple co-morbidities and risk factors for renal disease in the future, and interventions to limit this require consideration. Tubular disorders are common, and the full spectrum of renal disorders can occur. Proteinuria is particularly common in addition to renal calculi and effects of chelator therapy.
Monitoring or renal function in thalassaemia patients with renal insufficiency
Extensive monitoring of patients with thalassaemia in terms of the heart, endocrine organs (thyroid, testes, ovaries and pancreas) and the liver is routine practice, and the increased awareness of the kidney and an understanding of the potential mechanism of renal injury have lead to increased monitoring of the kidneys, especially in chelated patients.
Particular attention should be given to monitoring renal function with serum creatinine and eGFR and urinary protein creatinine ratio in patients who are at an increased risk of complications (patients with diabetes), have pre-existing renal conditions (or an eGFR < 60mL/min/1.73m2 and proteinuria), are elderly, have co-morbidities or are on medications, which may affect renal function. Ideally, 3 monthly serum creatinine/eGFR measurements (eGRF in preference and monthly in patients exposed to potential nephrotoxic agents), a simple urine dipstick to detect non-visible blood and a protein/creatinine ratio (in a morning urine sample), should be performed. More frequent measurements (monthly) in chelated patients should be performed based on current guidelines. Tubular function assessment using beta-2 microglobulin, urinary phosphate excretion and N-acetyl-B-d-glucosaminidase may be useful but referral to a nephrologist is perhaps advised, especially if haematoproteinuria is present.
Managing and preventing potential renal toxicity when using chelation therapy
Choice of iron chelator therapy should be individualised for each patient according to requirements and the clinical situation . Indeed, emerging data on combination therapy (deferiprone and deferoxamine or deferiprone and deferasirox) have demonstrated improved cardiac outcomes and reversal of endocrinopathies [95-97], and better efficacy and tolerability. Balocco et al.  demonstrated alternate use of deferasirox and deferiprone allowed tolerability with no adverse effects and efficacy in reducing serum ferritin. The future solution is perhaps the use of chelators flexibly, alternating use between chelators (i.e. drug free holidays) or at lower doses in combination to limit potential adverse effects as appears to be occurring in some centres and indeed in North America, combination oral chelation is a common practice. Chelator avoidance in high-risk groups (patients with diabetic, established renal insufficiency, significant proteinuria) remains debatable and a risk benefit consideration is necessary.