Cardiorenal syndrome in thalassemia patients

Background Cardiorenal syndrome (CRS), a serious condition with high morbidity and mortality, is characterized by the coexistence of cardiac abnormality and renal dysfunction. There is limited information about CRS in association thalassemia. This study aimed to investigate the prevalence of CRS in thalassemia patients and also associated risk factors. Methods Thalassemia patients who attended the out-patient clinic of a tertiary care university hospital from October 2016 to September 2017 were enrolled onto this cross-sectional study. Clinical and laboratory findings from 2 consecutive visits, 3 months apart, were assessed. The criteria for diagnosis of CRS was based on a system proposed by Ronco and McCullough. Cardiac abnormalities are assessed by clinical presentation, establishment of acute or chronic heart failure using definitions from 2016 ESC guidelines or from structural abnormalities shown in an echocardiogram. Renal dysfunction was defined as chronic kidney disease according to the 2012 KDIGO guidelines. Results Out of 90 thalassemia patients, 25 (27.8%) had CRS. The multivariable analysis showed a significant association between CRS and extramedullary hematopoiesis (EMH) (odds ratio (OR) 20.55, p = 0.016); thalassemia type [β0/βE vs β0/β0 thalassemia (OR 0.005, p = 0.002)]; pulmonary hypertension (OR 178.1, p = 0.001); elevated serum NT-proBNP (OR 1.028, p = 0.022), and elevated 24-h urine magnesium (OR 1.913, p = 0.016). There was no association found between CRS and frequency of blood transfusion, serum ferritin, liver iron concentration, cardiac T2*, type of iron chelating agents, or urine neutrophil gelatinase-associated lipocalin level. Conclusions CRS is relatively common in thalassemia patients. Its occurrence is associated with laboratory parameters which are easily measured in clinical practice.


Background
Thalassemia, characterized by a mutation of a globin gene, is a common cause of congenital chronic hemolytic anemia in many parts of Southeast Asia including Thailand [1]. The severity of anemia varies by mutation type and the need for regular transfusion. The latter categorizes patients into two groups, transfusion dependent thalassemia (TDT) and non-transfusion dependent thalassemia (NTDT) [2][3][4]. The major cause of death in thalassemia is heart failure secondary to chronic iron overload, a condition known as iron overload cardiomyopathy. In the past, thalassemia patients with iron overload cardiomyopathy usually died within the 1 st or 2 nd decade of life [5][6][7][8]. However, an improvement of medical care and iron chelation therapy has led to a decrease in cardiac death and improved life expectancy in these patients [7][8][9]. Nevertheless, such patients are still at high risk of complications associated with chronic hemolytic anemia, especially those associated with heart and renal abnormalities.
Heart abnormalities are still mainly caused by chronic iron deposition in myocytes which results in an increase in oxidative stress, inducing myocyte injury, increased myocardial fibrosis and decreased cardiac contraction [10]. Renal abnormalities in cases of thalassemia manifest themselves through tubular dysfunction, glomerular dysfunction, hyperfiltration and renal stones caused by chronic hemolytic anemia and chelation therapy [11][12][13][14][15][16][17]. A coexistence of cardiac and renal abnormalities is known as 'Cardiorenal syndrome (CRS) '. CRS has been defined by Ronco et al. [18] as abnormalities of heart and kidney which can be categorized into five types. Types 1 to 4 CRS, known as primary CRS, are caused by either a diseased heart or kidney whilst type 5 CRS, or secondary CRS, is defined as systemic conditions leading to simultaneous injury and/or dysfunction of both heart and kidney. Thalassemia is classified as a systemic disease so the simultaneous occurrence of cardiac and renal dysfunction is categorized as secondary, type 5 CRS. However, if the onset of cardiac and renal abnormalities differs, it could be categorized as any type of primary CRS (type [1][2][3][4]. There is currently little information concerning CRS in thalassemia. Therefore, this study aimed to determine the prevalence of secondary CRS in thalassemia patients and its associated risk factors.

Patient population
This was a cross-sectional cohort study. All thalassemia patients, including a TDT and NTDT population, from the hematology out-patient clinic at Chiang Mai University Hospital were enrolled from October 2016 to September 2017. Inclusion criteria were to have an underlying condition of thalassemia and be aged > 18 years. Exclusion criteria were congenital renal or heart disease and patients who had missing or inadequate 24h urine samples. Baseline characteristic data were collected. These included age, sex, thalassemia type, transfusion status, blood pressure, past medical history (in particular a history of diabetes, hypertension, dyslipidemia, thyroid and adrenal function), current medication and chelation therapy, iron status, previous cardiac T2* magnetic resonance imaging (MRI) and MRI for liver iron concentration (LIC), clinical history of heart failure, echocardiogram within 2 years (parameters including left ventricular ejection fraction, right ventricular systolic pressure, mean pulmonary arterial pressure, tricuspid velocity, diastolic dysfunction, E/A and E/E' ratio of mitral valve) and any other complications. Clinical and laboratory assessments were performed on 2 consecutive visits which were at least 3 months apart. Laboratory tests included serum NT-proBNP, spot urine protein, spot urine albumin, spot urine electrolytes including calcium, 24-h urine protein, and 24-h urine electrolytes. This study was approved by the ethical review board of the Faculty of Medicine, Chiang Mai University (STUDY CODE: MED-2559-043461 Research ID: 4346, Approval number 046/2017). All the patients provided written informed consent.

Definition and measurements
At the first visit, urine samples were collected from each patient for spot urine protein, albumin, creatinine, electrolytes and urinalysis. Blood samples were collected to test complete blood count (CBC), blood urea nitrogen (BUN), serum creatinine (SCr), electrolytes, ferritin, liver function test, serum NT-proBNP, lactate dehydrogenase (LDH), thyroid function test (TFT) and serum morning cortisol. A chest X-ray and electrocardiogram were also performed in all patients.
At the second visit, which took place at least 3 months after the first, the same lab tests were repeated and in addition 24-h urine samples were collected to measure 24-h urine protein, creatinine, and electrolytes. The sufficiency of 24-h urine collection was checked using urinary creatinine (Cr), sex and body weight and the following equations: If the ratio of Cr to body weight (BW) was < 10.8 or > 25.2 and a total urine volume < 1000 mL/d with a urinary creatinine level < 5 mmol/d, the urine collection was deemed incomplete [19,20].
We chose a follow up of at least 3 months to allow certainty as regards the definition and diagnostic criteria of chronic kidney disease. Estimated glomerular filtration rate (eGFR) was calculated using the CKD-EPI formula.
Kidney tubular injury was assessed by measuring urine NGAL. Spot urine samples (3 ml) from both visits of every patient were also stored at -80°C to enable urine neutrophil gelatinase-associated lipocalin (NGAL) assessment. When all samples had been collected the measurement of urine NGAL levels were performed using a chemiluminescent microparticle immunoassay (ARCHITECT Urine NGAL assay, Longford, Ireland). Comparison of the automated ARCHITECT assay and a manual ELISA assay can be used to indicate a good correlation between the methods (Spearman's rank correlation coefficient = 0.99) with a least squares linear regression line ARCHITECT = 0.93 (ELISA) + 4.2 (95% confidence interval for slope and intercept, 0.91 to 0.95 and − 0.8 to 9.2, respectively). The lower limit for detection NGAL of this assay is 1 ng/mL [21]. Chronic excretion of urine NGAL was defined as urine NGAL levels > 5 ng/ml at both visits.
NTDT is defined as thalassemia disease that does not require regular transfusion for survival. However, the definition of regular transfusion varies between studies. The criteria for NTDT used in this study were no more than three transfusions or 7 units of red cells in a year. TDT was defined as thalassemia that requires more than 3 transfusions a year and a transfusion free period of not more than 8 weeks and/ or the number of red cell transfusions in the past year was greater than 7 units [4].
Extramedullary hematopoiesis (EMH) was diagnosed from chest X-ray and/or CT-scan and/or MRI.
In this study, CRS [18] was defined as the following: 1. Heart failure and/or cardiac abnormalities. The diagnostic criteria used to assess heart failure in this study were based on the 2016 European Society of Cardiology (ESC) Guidelines for the diagnosis and treatment of acute and chronic heart failure [22]. These criteria classify patients into heart failure with a preserved ejection fraction (HFpEF), midrange ejection fraction (HFmrEF) and reduced ejection fraction (HFrEF) as shown in supplementary table S1. Typical symptoms of heart failure include breathlessness, orthopnea, paroxysmal nocturnal dyspnea, reduced exercise tolerance, fatigue, tiredness, increased time to recover after exercise, and ankle swelling. Typical signs of heart failure include elevated jugular venous pressure, hepatojugular reflux, third heart sound (gallop rhythm), and laterally displaced apical impulse. Cardiac abnormalities included structural remodeling identified by echocardiogram (such as left ventricular hypertrophy, cardiomegaly and iron overload cardiomyopathy or hemochromatosis) and cardiac dysfunction. Diagnostic criteria for left ventricular hypertrophy were based on QRS voltage criteria i.e. R wave in V5/V6 plus S wave in V1/V2 exceeds 35 mm in height (SV 1-2 + RV 5-6 > 35 mm). Diagnostic criteria for right ventricular hypertrophy were right axis deviation with tall R-waves in RV leads and deep S-waves in LV leads. Left atrial abnormality was defined as P-wave > 120 millisecond and wide notched P-wave > 40 milliseconds. Right atrial abnormality was defined as upright P-wave in lead II > 2.5 mm [23]. The criterion for diagnosis of cardiomegaly from a standard chest X-ray was a cardiothoracic ratio > 0.5 [24]. In accordance with Thalassemia International Federation guidelines, we defined hemochromatosis using the cardiac T2* technique and magnetic resonance imaging of the liver to determine liver iron concentration (LIC). Hemochromatosis was diagnosed when cardiac T2* was less than 20 milliseconds or the LIC more than 7 mgFe/g dry weight [3,4].

Chronic kidney disease. The KDIGO 2012 Clinical
Practice Guidelines for the Evaluation and Management of Chronic Kidney Disease [25] state that the condition is diagnosed if abnormalities of kidney structure or function are present for > 3 months. These include: -GFR < 60 ml/min/1.73 m 2 or -Presence of 1 or more markers of kidney damage: albuminuria (Albumin excretion rate (AER) ≥30 mg/ 24 h; Albumin to creatinine ratio (ACR) ≥30 mg/g [≥3 mg/mmol]); urine sediment abnormalities; electrolyte and other abnormalities due to tubular disorders; abnormalities detected by histology; structural abnormalities, kidney stones detected by imaging; history of kidney transplantation.

Statistical analysis
Statistical analysis was carried out using SPSS version 23.0. Sample size was calculated by estimation for a single proportion model. Previously published data on prevalence of renal and cardiac abnormalities was used to calculate sample size [26][27][28]. Estimated sample size was 90 for an alpha of 0.05 and power 0.8. Pearson's Chi-square or Fisher's exact test was performed to calculate the association between CRS and other sets of categorized data and to calculate odds ratios. Normality was checked using Kolmogorov-Smirnov and Shapiro-Wilk methods. The statistical significance of differences in continuous data sets was calculated using an independent sample t-test. Mann-Whitney U-test was used for non-normally distributed variables. Multivariable analysis was performed using a binary logisticregression model which included all variables from the univariate logistic regression with significant differences. A p value of < 0.05 was considered significant. The univariable analysis indicated that CRS was more frequently observed in β 0 /β 0 thalassemia when compared with other types of thalassemia (p = 0.04) ( Table 1). The detail of thalassemia type is provided in Table 2. Thalassemia with secondary CRS also showed a significant association with pulmonary hypertension (PHT) (p < 0.001), hemochromatosis (p = 0.047) and the presence of extramedullary hematopoiesis (EMH) (p = 0.02). In addition, patients with CRS had a significantly higher level of serum NT-proBNP, 24-h urine protein and 24-h urine magnesium than those without CRS.

Results
Patients with chronic excretion of urine NGAL at a level higher than 5 ng/ml showed a significant association with the occurrence of CRS in the univariable analysis (OR 2.82, p = 0.038) ( Table 3), but not in the multivariable analysis (Table 4). Chronic excretion of urine NGAL > 5 ng/ml also showed a significant association with combined deferoxamine and deferiprone treatment (OR 3.69, p = 0.011), female gender (OR 5.15, p < 0.001), hemochromatosis (OR 3.01, p = 0.012), elevated serum LDH (OR 2.81, p = 0.018) and chronic proteinuria (OR 2.63, p = 0.025). The 24-h urine protein level was significantly higher in patients with chronic excretion of urine NGAL > 5 ng/ml than those without (p = 0.035).
A binary logistic-regression model was performed in the case of the multivariable analysis and confirmed an independent association between CRS with the type of thalassemia [β 0 /β E VS. β 0 /β 0 thalassemia (OR = 0.005, p = 0.002)], presence of EMH (OR 20.549, p = 0.016), presence of PHT (OR 25.455, p = 0.016), elevated serum NT-proBNP (OR 1.028, p = 0.022) and elevated 24-h urine magnesium (OR 1.913, p = 0.016) as shown in Table 5. There was no association between CRS and sex, age, splenectomy, transfusion status, mean hemoglobin level, mean serum ferritin, liver iron concentration, serum LDH, thyroid function, serum cortisol, hypogonadism, all echocardiogram parameters, cardiac T2* and type of iron chelation.
Thalassemia is a highly prevalent condition in countries of Southeast Asia including Thailand, and currently there is restricted data regarding the coexistence of renal and cardiac abnormalities or CRS with the condition. This study has demonstrated that the prevalence of CRS in thalassemia is 27.8%. To the best of our knowledge, our study was the first study that has demonstrated the prevalence of CRS in thalassemia. The occurrence of CRS showed an association with the type of thalassemia (β 0 /β 0 thalassemia), EMH, PHT, increased 24-h urine magnesium and elevated serum NT-proBNP. However, the cut points for BNP and NT-proBNP in our study were below those set for those with CKD and heart failure in previous studies [38,39]. An association between CRS and EMH/ PHT may be explained by the presence of chronic severe anemia with inadequate transfusion which resulted in cardiac and renal hypoxia and dysfunction manifested by elevated serum NT-proBNP levels and magnesiumuria. Chronic anemia-associated systemic hypoxia generally enhances cardiac compensation by initially increasing cardiac output which ultimately leads to pathologic cardiac remodeling such as chamber enlargement if the anemia is not corrected. Chronic activation of the reninangiotensin-aldosterone system (RAAS) may be a major key to treating secondary CRS in thalassemia patients since cardiac and renal hypoxia generally stimulates RAAS. It is noteworthy that persistent RAAS activation can induce myocardial fibrosis, renal tubular damage, efferent arteriole constriction causing glomerular hypertension, proteinuria and renal fibrosis, thereby leading to CRS [40].
The degree of tissue hypoxia may explain why CRS showed an association with β 0 /β 0 thalassemia. β 0 /β 0 thalassemia is caused by total deletion of the beta globin gene resulting in a severe misbalance of the alpha and beta globin chains and severe anemia. β 0 /β E thalassemia sufferers however, have less severe anemia than β 0 /β 0 thalassemia because some beta globin chains, despite being a defective form, are still produced. Although hemoglobin E is an abnormal form of hemoglobin and made of abnormal globin chains, it can still carry and deliver oxygen to tissues.
Neutrophil gelatinase-associated lipocalin (NGAL) is a 25 kDA, low molecular weight protein, majorly found in neutrophil granules and renal tubular cells. NGAL is involved in iron transportation, iron binding and renal cell repair. It passes freely through the glomeruli then is for the most part reabsorbed in the proximal renal tubules. NGAL can be elevated in inflammatory conditions and renal diseases such as: autosomal dominant polycystic kidney disease; immunoglobulin A nephropathy; HIV nephropathy; contrast-induced nephropathy; urinary tract infections and renal tubular injury. Elevation of urine NGAL is associated with decreased estimated GFR, urinary albumin excretion, increased serum NT-proBNP levels and may be potentially used for early detection of AKI [41,42]. However, several studies have shown inconclusive results as regards urine NGAL as a predictor of CKD progression [43][44][45][46][47][48].
Clinical studies in cardiac surgery settings, after contrast infusion, critical illness, and traumatic patients, have demonstrated an association between NGAL and CRS Cardiorenal syndrome, OR odd ratio, EMH extramedullary hematopoiesis, yrs years, ms millisecond early detection of AKI. Both urine and serum NGAL have been shown to be elevated preceding the elevation of serum creatinine in AKI patients. Elevation of urine or serum NGAL has also been observed in CRS patients, especially in cases of type 1 CRS, and may have been beneficial in the diagnosis of other types [18,[49][50][51][52][53]. However, recent large prospective cohort study AKINES IS has shown serum and urine NGAL was not superior to serum creatinine for predict worsening renal failure or need for renal-replacement therapy in patients with acute heart failure (CRS type 1) [54][55][56].
A study of relevant literature shows the cutoff level point for the level of urine NGAL for AKI or CKD progression varies from 10 to 500 ng/ml depending on patient population and conditions but to date there is no generalized standard cutoff for urine NGAL [21,41,42,49,57]. In our study, the univariate analysis indicated that a chronic urine NGAL level higher than 5 ng/ml showed an association with female gender, combined deferoxamine (DFO) and deferiprone (DFP) treatment, hemochromatosis, elevated serum LDH, chronic proteinuria, and increased 24-h urine protein. Proteinuria in thalassemia may be contributed to by glomerular or tubular injury, but evidence from our study supports tubular injury to be the main cause as urine NGAL was a tubular damage marker and was not elevated in cases of glomerular injury [58,59]. The cut-off of urine NGAL in this study quite low compared to other studies and we hypothesized that thalassemia patients had glomerular hyperfiltration which can lower urine NGAL causing the low cut-off. The association between hemochromatosis and urine NGAL excretion may be due to iron toxicity that produces free radicals to injure renal tubular cells. Combined use of iron chelators, DFO and DFP, results in a greater severity of iron overload and renal tubular injury. This may explain its association with chronic urine NGAL excretion. In addition, both DFO and DFP may be directly toxic to renal tubular cells as they synergistically enhance iron excretion via the kidney. This therapy also results in increased iron accumulation in renal tubular cells which causes tubular cell damage [13,60]. In this study we found that the threshold of > 5 ng/mL was associated with CRS. However the threshold of urine NGAL in our study was quite low compared to other studies [21,61]. As previously mentioned, the cut point of urine NGAL can vary depending on patient population. But all studies were in agreement that NGAL could be a useful marker for predicting tubular damage and AKI [60].
A limitation of this study is that it is a cross-sectional study with only a 3 month follow-up. Prospective long term follow-up would quite possibly detect more detailed dynamics of the disease and other complications, in particular those related to renal abnormalities and changes in eGFR. Moreover, a prospective long term study or a study with increased follow-up times and shorter intervals may be useful in pinpointing more precise onset of cardiac and renal abnormalities in order to indicate a specific type of CRS. In this study, data collection of renal abnormalities was based on laboratory criteria while renal symptoms may have not developed at that point. Echocardiographic data in our study also had some missing data. The majority of cases of renal abnormalities in this study were characterized by proteinuria which is common in type 2 CRS, and can be found in types 4 and 5 CRS. Apart from conventional renal markers such as serum creatinine and eGFR, the tests for which are not sensitive enough for early detection of renal injury, only a single novel kidney injury biomarker, urine NGAL, was used in the present study. Using a panel of biomarkers may give a higher level of sensitivity for the detection of renal abnormalities. One interesting alternative marker which provides an early indication of renal damage is 'renal functional reserve' as some thalassemia patients tend to have glomerular hyperfiltration which may reflect a functional reserve response to stress. Chronic stress to the kidney could result in low renal functional reserve despite a normal eGFR or serum creatinine level [62]. However, the cut-off of urine NGAL in this study was quite low compared to levels described in previous studies. These may have caused false positive readings of elevated urine NGAL. Urine NGAL as a standalone measure is not a precise indicator for a diagnosis of kidney injury. The inclusion of other markers such as TIMP-2, KIM-1 in addition to NGAL in the future could improve the precision. Renal ultrasound was not routinely done in all patients unless there was a clinical indication/suspicion which may have resulted in structural diseases of the kidneys and urinary tract as well as kidney stones being missed.
There were also two important confounding factors in this study into CRS in thalassemic patients. First, iron chelation therapy especially deferasirox can cause renal failure and proteinuria. However, this factor would be unlikely to impact on the findings of our study since only one patient had an eGFR below 60 ml/min/1.73 m 2 . Second, heart failure symptoms can mimic anemic symptoms, but to prevent this potential misinterpretation we used physical examination to ensure the diagnoses were accurate. We also recorded symptoms after blood transfusion, which should be improved if the patient had symptoms indicating anemia.

Conclusions
CRS is relatively common in thalassemia patients, especially in cases of β 0 /β 0 thalassemia. This condition is associated with the presence of EMH and PHT. Elevated serum NT-proBNP and magnesiumuria may be useful markers for the detection of CRS in thalassemia cases.