Role of non-invasive assessment in prediction of preclinical cardiac affection in multi-transfused thalassaemia major patients

Abstract

Background

The principal cause of mortality and morbidity in β-thalassemia major (β-TM) is the iron overload as these patients receive about 20 times the normal intake of iron, which leads to iron accumulation and damage in the liver, heart, and endocrine organs. Chronically transfused patients used to die from cardiac iron overload in their teens and twenties. Monitoring of iron status through cardiac magnetic resonance imaging (CMRI) has replaced the conventional methods of assessment, yet this modality is not readily available in centers where the disease distribution is maximal.

Objectives

The aim of this work is to study some simple non-invasive tools and their abilities to predict preclinical cardiac affection reflecting cardiac iron deposition (CID) in multi-transfused β-TM patients taking the T2* CMRI as a gold standard reference test.

Methods

One hundred consecutive multi-transfused, clinically stable β-TM patients with age ranging from 16 to 30 years (mean ± SD, 21.1 ± 3.9) were included in this study. Assessment of serum ferritin, serum hepcidin, and high-sensitivity C-reactive protein as well as cardiac assessment by echo-doppler and 24-hour Holter were used to predict CID, and consequently predict preclinical cardiac affection, in reference to CMRI results as the standard method of cardiac iron assessment.

Results

According to CMRI results, patients were subdivided into a group of 42 patients with detectable myocardial iron (T*≤ 20 ms) and a group of 58 patients with no detectable myocardial iron (T* > 20 ms). No differences in age, gender, or distribution of splenectomized patients were observed between both groups. Patients with detectable myocardial iron received significantly higher number of transfusions per year than those with no detectable myocardial iron (mean ± SD, 14.6 ± 1.7 vs. 12.5 ± 1.7; P < 0.001) yet comparable levels of serum ferritin, serum hepcidin, and hepcidin/ferritin ratio (P > 0.05) were noted. Cardiac iron detection was associated with significantly lower heart rate (mean ± SD, 75 ± 6.1 vs. 80 ± 6.9; P < 0.001), lower left ventricular ejection fraction (LVEF) (mean ± SD 60.1 ± 3.2 vs. 70.1 ± 2.8; P < 0.001), and higher total number of premature ventricular contractions (PVCs) (median 78 vs. 14; P < 0.001). The group with CID comprised significantly more patients with left ventricular diastolic dysfunction (15/42, 35.7% vs. 3/58, 5.2%; P < 0.001); PVCs ≥10/hour (13/42, 31% vs. 2/58, 3.4%; P < 0.001); episodes of sinus pauses (6/42, 14.3% vs. 1/58, 1.7%; P < 0.05); episodes of high-grade atrio-ventricular block (5/42, 11.9% vs. 1/58, 1.7%; P < 0.05) compared to the group with no (CID). In presence of normal LVEF, detection of 10 or more PVCs per hour was the most predictive of cardiac iron loading with a positive predictive value of 86.7% and specificity of 96.6%, and the highest likelihood ratio (9.09). Detection of more than 22 PVCs/24 hours had the best sensitivity (81%) and the best negative predictive value (84%). The positive likelihood ratio of the studied parameters was highest in case of presence of PVCs ≥10/hour and lowest in case of average heart rate with a cut-off level of ≤77.5 bpm (9.09 and 1.46, respectively).

Conclusion

Our results support our hypothesis that monitoring β-TM patients with echo and Holter electrocardiogram can help in the detection of preclinical cardiac affection in centers lacking cardiac MRI; however, due to relatively low sensitivity they can not fully replace CMRI. Further work is needed to identify additional simple parameters that can form a diagnostic model with adequate sensitivity.

Keywords:

• Preclinical cardiac affection
• B thalassaemia major
• Transfusion
• Cardiac iron deposition
• Hepcidin
• Cardiac MRI

Introduction

Beta-thalassemia is an inherited hemoglobin disorder resulting from either homozygous or double heterozygous inheritance of two abnormal genes from the β-globin locus, leading to defective synthesis of the β-globin chain with the resultant chronic dyserythropoietic anemia and subsequently hemolytic anemia.¹ The principal cause of mortality and morbidity in β-thalassemia is the iron overload. In the homozygous form, β-thalassemia major (β-TM), this is mainly attributed to the repeated transfusions required for the treatment of the disease, and to a lesser extent to the increased iron absorption as a result of the strong suppressive effect of high erythropoietic activity on the expression of hepcidin, the main circulating iron hemostatic peptide.^2,3 Consequently, β-TM patients receive about 20 times the normal intake of iron, which leads to iron accumulation and damage in the liver, heart, and endocrine organs.⁴

Chronically transfused patients used to die from cardiac iron overload in their teens and twenties.⁵ Advances in the treatment modalities of β-TM have led to improved prognosis, and nowadays, more patients survive up to the third or fourth decade.⁶ Although improvements in survival have been achieved, iron-induced cardiotoxicity remains a leading cause of morbidity and mortality among those patients.⁶ Despite the fact that cardiac function remains normal for many years, once heart failure symptoms become evident, death usually occurs within 1 year.^7,8

Iron overload-related cardiomyopathy is reversible, but the diagnosis is often hindered by the late appearance of symptoms and echocardiographic abnormalities.⁹ Indeed assessment of left ventricle (LV) structure and systolic function in this clinical setting have limitations namely, β-TM patients have growth retardation with a small body surface area compared with unaffected subjects; the basal high cardiac output seen in chronic anemia can mask the ventricular dysfunction; and the degree of chronic anemia may affect the ventricular size. Therefore, both definition and interpretation of LV size and function in β-TM are variable and somewhat conflicting in the available literature.¹⁰

Despite being frequently used to monitor iron overload, serum ferritin is a labile marker of iron balance since it varies with inflammation, ascorbate status, and intensity of transfusion therapy.¹¹ The invasiveness of liver iron measurement by biopsy limited its use as a routine screening at most institutions;¹¹ moreover, liver iron deposition was shown to be non-correlating with myocardial iron deposition.¹² Endomyocardial biopsy is also unreliable for measuring myocardial iron because of invasiveness, sampling error with very small biopsy samples.^13,14

The need for an alternative non-invasive measurement of myocardial iron led to the development of an optimized cardiac T2* magnetic resonance imaging (MRI) technique. Myocardial T2* is an easily quantifiable, clinically robust, and highly reproducible measurement technique.¹⁵ It is, now, the gold standard for assessing myocardial iron, yet the very high cost and scarce availability in tertiary centers limited its use as a screening tool.⁹ The recognition that cardiomyopathy and arrhythmia are the most frequent causes of death in thalassemia¹⁶ has provided further stimulus for the development of additional non-invasive screening tools for the cardiac iron overload and toxicity.

In our work we planned to study some non-invasive tools to predict early cardiac iron deposition (CID) in multi transfused thalassemia major patients in comparison with the T2* cardiac MRI (CMRI) as a gold standard test.

Subjects and methods

Between June 2010 and May 2012, 100 consecutive multi-transfused clinically stable β-TM patients either homozygous or compound heterozygous requiring regular blood transfusion therapy from the age of 2 years were included in our study. Patients were 16 years or older, following in two centers in eastern province of Saudi Arabia. Clinical stability was defined as patients with no acute anemia related complications, no clinical heart failure or cardiac medications on chart. Patients were considered to be multi- transfused if they received more than 10 transfusions per year.

All enrolled patients gave informed consent. The study was reviewed and accepted by the ethical and research committee of both institutes. The management protocol for thalassemia patients and indication, frequency, and amount of blood transfusion were according to the international published guidelines.¹⁷

All patients had a full clinical examination after a thorough medical history. Patients were free of heart failure on clinical examination, with normal cardiac silhouette on chest radiography, and normal LV systolic function on echocardiogram. We excluded patients with arterial hypertension or other causes of diastolic dysfunction, valvular or congenital heart disease, and Holter recordings lasting less than 16 hours or with technical deficiencies resulting in unreliable analysis. Patients with MRI incompatible metallic implants (metallic valves, joints, or plates) or claustrophobia that may preclude the MRI were also excluded.

The design of the work at hand was cross section observational study to be carried out in a single visit 5 to 10 days from the last transfusion. Patients with a delay between the MRI, laboratory, echocardiography (ECHO), and/or Holter of more than 15 days were excluded from the study.

Laboratory protocol

For laboratory investigations, ethylenediaminetetraacetic acid anti-coagulated samples were withdrawn for a standard blood count (Cell-Dyn Sapphire; Abbott Laboratories, IL, USA). Serum from clotted samples was used for determination of ferritin levels using an immunoluminescent assay (Vitros 5.1; Ortho Clinical Diagnostics, Inc, Buckinghamshire, UK) and high-sensitivity C-reactive protein (hs-CRP) using an immunoturbidimetric assay (Vitros 5.1; Ortho Clinical Diagnostics, Inc, Buckinghamshire, UK). Hs-CRP was considered positive if levels were higher than 1.0 g/l according to American Heart Association Classification.¹⁸

For serum hepcidin level determination, samples on plain vacutainer tubes were allowed to stand for 2 hours, followed by centrifugation at 1500g for 15 minutes and serum was collected and stored at −80°C for further analysis using a commercially available competitive enzyme-linked immunoassay (c-ELISA) (Bachem, St Helens, Merseyside, UK). Results were determined from standard curves developed from calibrators run simultaneously with study samples.

Echocardiography

M-mode; two-dimensional; color Doppler flow imaging with pulsed and continuous wave spectral analysis of transvalvular flow; were performed for all patients using the iE33 Ultrasound System (Philips Medical Systems, Andover, MA, USA). Appropriate transthoracic transducer frequency selection was made at the time of the evaluation. Cardiac measurements were performed according to the guidelines of American Society of Echocardiography.¹⁹ Patients with left ventricular ejection fraction (EF) <55% were excluded from the study. Trans-valvular Doppler flow analysis was used to assess left ventricular diastolic functions whereas inversion of the E/A wave ratio (E/A ratio <1) of the trans-mitral flow was considered as an index of left ventricular diastolic dysfunction (LVDD).

Twenty-four-hour ambulatory Holter monitoring

All patients underwent 24-hour Holter monitoring using three channel real-time Digital Holter Monitoring Recorders (HILLMED HM–HOLTER 600 LPII recorder, HILLMED Corporation, Miami, USA) and monitoring of the bipolar leads CM5 & CM3 and modified aVF.

Holter recordings were analyzed using the (HILLMED HM–HOLTER device) on a Microsoft compatible personal computer. The average heart rate, total number of premature ventricular contractions (PVCs), presence of PVCs ≥ 10/hour, episodes of non-sustained ventricular tachycardia (NSVT) (defined as ≥ 3 consecutive PVCs with a rate ≥ 100 beats/minute), high grade (Mobitz II second degree and third degree heart block), and sinus pauses of greater than 2.5 seconds were obtained for each patient.

Cardiac magnetic resonance imaging

All patients underwent MRI examination in the supine position with electrocardiogram (ECG) and breath follow-up pad using a General Electric 1.5-T CVi system (General Electric Medical Systems; TEs of 5.6, 6.5, 7.5, 9, 12, and 15 ms). For T2* evaluation, images were taken from the short axis mid-ventricular line, using black blood dual-echo cardiac triggered turbo field echo (TFE) sequence, by using two echo times (TE). A single-shot dual-echo T2*-weighted short axis fast-field-echo sequence TR = 12 ms; flip angle 30°; FOV, 320 mm; RFOV, 100%; slice thickness 10 mm) was acquired in all patients with the following echo times: TE1 = 4.6 ms; TE2 = 9.2 ms. For each patient, signal intensity of three small regions of interest placed in basal, middle, and apical inter-ventricular septum was measured. T2* was calculated using the following formula: T2* = −ΔTE/ln(SITE2/SITE1). A mean myocardial T2* value was calculated for each patient by making use of the T2* values of the three regions of interest. T2* ≤ 20 ms was considered as a sign of CID.^11,20

Statistical analysis

Continuous variables were expressed in the form of range, mean ± SD or median and compared with the Student t test when parametric or Mann–Whitney test when non-parametric. Categorial variables were expressed in the form of numbers and percent and comparisons were made by χ² or Fisher's exact test as appropriate. All tests being 2-tailed and a P value of <0.05 was considered statistically significant. Analysis was made by Graphpad prism software version 4.0 (Graphpad software Inc, La Joila, CA, USA).

The mean and median levels were used as cut-off values for parametric and non-parametric variables respectively. Positive predictive value (PPV) was calculated according to the formula [true positive/(true positive + false positive)] [TP/(TP + FP)] and negative predictive value (NPV) was calculated by the formula [true negative/(true negative + false negative)] [TN/(TN + FN)]. Sensitivity was determined by dividing TP/(TP + FN) and Specificity was calculated by dividing TN/(FP + TN).²¹

We used likelihood ratios for assessing the value of performing the tests we examined. We used the sensitivity and specificity of the test to determine whether the result is useful in predicting preclinical cardiac affection. The likelihood positive ratio of a test was determined using the equation: LR + =Sensitivity/(1–Specificity).²²

Results

This cross-sectional observational study was carried out on 100 consecutive clinically stable β-TM patients, following in The Royal Commission Hospital, and King Abdulaziz Naval Hospital in the eastern province of Saudi Arabia during the period from June 2010 to May 2012.

Forty-eight of our study population were male (48%), our study population ranged in age from 16 to 30 years with a mean of 21.1 ± 3.9 years. All of them were diagnosed as having β-TM and managed according to the international published guidelines for diagnosis and management of β-TM. All patients were clinically stable with no acute anemia related complications, no clinical heart failure or cardiac medications on file. Thirty-six patients (36%) were splenectomized, 13 patients (13%) had positive hs-CRP, 18 patients (18%) had LVDD, 15 patients (15%) had PVCs ≥ 10/hour as detected by Holter recording. Two patients (2%) had recordable NSVT, six patients (6%) had episodes of high-grade atrio-ventricular (AV) block, and 7 patients (7%) had one or more episodes of sinus pauses. The remaining demographic, clinical characteristics, ECHO, Holter, and laboratory results of the study group are summarized in Table 1.

Table 1. The demographic, clinical characteristics, ECHO, Holter, and laboratory results of the study group

We classified our study population according to the results of CMRI into two groups; those with detectable myocardial iron having a T2* of 20 ms or less (T* ≤ 20 ms) representing the group of preclinical cardiac affection (42 patients) and those with no detectable myocardial iron having a T2* more than 20 ms (T* > 20 ms) representing the group with no preclinical cardiac affection (58 patients).

The two studied groups were comparable in data regarding the age, gender, incidence of splenectomy, hemoglobin levels, and hs-CRP concentrations (Tables 2 and 3).

Table 2. Qualitative clinical data of β-TM patients in relation to detection of myocardial iron by MRI

Table 3. Quantitative clinical data of β-TM patients in relation to detection of myocardial iron by MRI

When comparing the two studied groups, the group with preclinical cardiac affection (T2* ≤ 20 ms) had significantly more patients with LVDD (35.7 vs. 5.2%, P <0.001), PVCs ≥ 10/hour (31 vs. 3.4%, P < 0.001), episodes of Sinus pauses (14.3 vs. 1.7%, P < 0.05) and episodes of high-grade AV Block (11.9 vs. 1.7%, P < 0.05). The two patients who had NSVT were both in the group with preclinical cardiac affection (T2* ≤ 20 ms) yet the difference from the group without preclinical cardiac affection did not reach statistical significance (P > 0.05) due to the small sample size. Differences in the remaining qualitative parameters are summarized in (Table 2).

Although patients with detectable myocardial iron (T2* ≤ 20 ms) had higher serum ferritin levels and lower hepcidin concentrations than those with no detectable myocardial iron (T* > 20 ms), the difference was not proved to be of statistical significance (P > 0.05) (Table 3). The hepcidin/ferritin ratio was significantly less than one in either group but with no statistically significant difference between them (P > 0.05) (Table 3).

Patients with preclinical cardiac affection (T2* ≤ 20 ms) had significantly higher number of transfusions/year (14.6 ± 1.7 vs. 12.5 ± 1.7, P < 0.001) and higher transfusion rates (183 ± 34.1 ml/kg/year vs. 161 ± 31.3, P = 0.001), lower average heart rate (75.0 vs. 80.0, P < 0.001), lower left ventricular ejection fraction (LVEF) (60.1 vs. 70.1, P < 0.001), and higher total number of PVCs (median 78 vs. 14, P < 0.001) when compared to patients without preclinical cardiac affection (T2* > 20 ms). (Table 3)

Prediction of preclinical cardiac affection

In presence of normal LVEF, transfusion frequency/year with a cutoff level of ≥13.5 was predictive for preclinical cardiac affection with sensitivity of 73.8% and specificity of 67.2%. Average heart rate with a cutoff level of ≤77.5 bpm was predictive for preclinical cardiac affection with sensitivity of 61.9 % and specificity of 58.6%, total number of PVCs/24 hours with a cutoff level of ≥22 beats was predictive for preclinical cardiac affection with sensitivity of 81.0 % and specificity of 72.4 %.

The PPV, NPV, sensitivity, and specificity for the detection of LVDD, PVCs ≥10/hour, sinus pauses >2.5 seconds, or high-grade AV block are summarized in Table 4.

Table 4. Predictors of preclinical cardiac affection

We found that, in presence of normal LVEF, detection of ten or more PVCs per hour was the most predictive of preclinical cardiac affection with a PPV of 86.7% and specificity of 96.6%, and the highest likelihood ratio (9.09). Sinus pauses of 2.5 seconds or more; high-grade AV block, and presence of LVDD share the same finding of good PPV with high specificity and low sensitivity. On the other hand, detection of more than 22 PVCs/24 hours has the best sensitivity and the best negative predictive test.

The positive likelihood ratio of the studied parameters was highest in case of presence of PVCs ≥10/hour to lowest in case of average heart rate with a cutoff level of ≤77.5 bpm (9.09, 1.46, respectively).

Discussion

Iron overload in β-TM occurs due to a combination of repeated blood transfusions, and excessive gastrointestinal absorption.¹⁶ β-TM has known complications of congestive heart failure, sudden cardiac death, and arrhythmias that are related in part to iron overload. Since the introduction of deferoxamine in the early 1970s, life expectancy has improved dramatically. However, despite improvements, mortality in middle age continued to be a problem¹⁰ due to several reasons; firstly, the difficulty of compliance to deferoxamine; secondly, the late onset and ominous nature of cardiac symptoms and imaging evidence of ventricular dysfunction;⁷ thirdly, the insidious occult cardiac iron accumulation suggested by development of cardiac symptoms despite apparently adequate control of hepatic iron stores; lastly, the reliance of many centers exclusively on serum ferritin to track somatic iron stores, which although remains an important monitoring tool, is a poor marker of iron balance due to influences on its levels by various factors.¹¹

Some centers routinely screen their patients with CMRI to detect preclinical cardiac affection. A cardiac T2* of 20 ms or less is indicative of CID and a T2* below 10 ms conveys significant prospective risk for cardiac dysfunction.^23,24 However, such a technique requires highly equipped and specialized centers, therefore, in our work we aimed to study some simple non-invasive measurements (cardiac and laboratory) in β-TM patients and their ability to predict cardiac iron overload in reference to the T2* CMRI as a gold standard test for assessment of preclinical cardiac affection in multi-transfused β-TM patients. Up to date, there is no clear definition of cardiac involvement in thalassaemia syndromes, with particular attention to β-TM. The most recent edition of Braunwald's Heart Disease²⁵ describes cardiac involvement in hemochromatosis as ‘a mixed dilated cardiomyopathy – restrictive cardiomyopathy pattern with both systolic and diastolic dysfunction often with associated arrhythmias’. We proposed the parameters at hand to be early indicators of cardiac iron overload and compared them to the standard published data from CMRI.

Our work was a cross-sectional observation study performed on 100 β-TM patients in two centers in eastern province KSA where this hereditary hemolytic anemia constitutes a significant health problem. Although it was considered a fact that β-TM is a major pediatric problem, treatment advances have led to prolonged patients' survival. The study was limited to the adult population (≥16 years) as it was demonstrated that at least 13 years of chronic transfusion therapy are required to produce cardiac iron loading²⁴ with no reports of detectable cardiac iron overload observed in children under the age of 9.5 years.²⁶ An additional technical reason is to limit the confounding effect of the body mass index on the ECHO parameters and to eliminate the effect of age on the other parameters studied in this research. The maximum age recruited was 30 years old, and none of our patients had overt cardiac manifestations. The life expectancy of thalassaemia patients was variable in the available literature and this variability was primarily a result of treatment advances and better health awareness in the developed world in contrast to under developed countries where the disease burden remains to be identified. Data from Modell's study, suggested that 50% of patients from Great Britain died before reaching the age of 35 years before introducing oral chelation therapies. This figure has dramatically changed by introducing MRI monitorization and oral chelators. A 71% reduction in annualized death rate from iron overload between 2000 and 2004 has been reported.⁶

No interference was made in the management protocol of each institute and all patients were treated by chelation therapy according to serum ferritin levels. The transfusion rates of the studied population ranged from 10 to 17 per year (mean, 13.4) and patients with less than 10 transfusions per year were excluded, as the main purpose is to investigate the effect of myocardial iron overload resulting from chronic repeated blood transfusion. The study was done in a single visit within a more or less fixed interval from the last transfusion (5–10 days), patients with excessive delay (till the next transfusion) in performing any of the research measurements were excluded, to eliminate the possible effects of fluctuating serum iron and ferritin level between the transfusions.

Patients were classified according to T* values into 42/100 having detectable myocardial iron and 58/100 with no such iron detection. Not surprisingly, patients with cardiac iron loading received a significantly higher number of transfusions per year; however, their higher serum ferritin level than those without detectable myocardial iron did not reach statistical significance. This particular point is still controversial as some published data report serum ferritin to be a non-reflective measure of iron overload in general and in particular CID,¹¹ while others demonstrate a negative correlation between its level and myocardial T2*.^9,10

Serum hepcidin levels did not significantly differ in both groups, which could be attributed to the timing of the sampling applied in this study (5–10 days from the last transfusion) as hepcidin levels in chronically transfused patients were shown to be dynamic and increase post-transfusion due to both increased iron load and the alleviation of the ineffective erythropoiesis.^27,28 The relative suppression of hepcidin concentration given the high degree of iron loading in the studied patients, as evidenced by the very low hepcidin/ferritin ratio, could be regarded as a contributing factor to iron overload other than repeated transfusions. However, the lack of relation of this ratio to cardiac iron loading in our data could imply that all patients were exhibiting similar characteristics of their iron homeostatic factors with their only difference being in the transfusion load. Rapid iron accumulation is known to be associated with prominent deposition of iron in the heart and some endocrine organs,³ which could account for the increased cardiac iron accompanying higher transfusion rates (and consequently higher rates of iron accumulation) noted in the present study, despite the comparable serum ferritin and hepcidin levels in all patients.

Although Detterich et al.²⁹ showed that hs-CRP, a non-specific marker of systemic inflammation, was higher in patients with cardiac iron overload, our patients did not demonstrate such an association, with only 13% of them having values >1 g/l, suggesting that chronic vascular inflammation may not be an essential part of β-TM.

In this study, patients with CMRI ≤20 ms were more likely to suffer from LVDD denoting that myocardial iron deposition affects the diastolic function of the heart prior to deterioration of the systolic function. Indeed, they had significantly lower, yet normal, EF than patients with no iron overload; findings which are similar to those observed in other works;^9,10 which could be explained by the absence of augmentation in the systolic function (high cardiac output status that accompany chronic anemia) in the iron overloaded hearts.

Arrhythmias detected by Holter, similar to those observed in the present study, have been well documented in published works regardless of chelation status.^30,31 Fifteen percent of our patients had at least one form of cardiac arrhythmias as detected by Holter recording. Ventricular tachycardia was recorded in 2% of our patients in comparison to slightly less than 1% in the Kirk et al.³² study population.

Both forms of tachyarrhythmia present in our study were more prevalent in patients with cardiac iron overload who had a significantly higher total PVCs count. This could be attributed to the repolarization abnormalities, such as ST and T wave abnormalities and the prolonged QT and QTc that are characteristics of iron-overloaded hearts.²⁹ The changes in repolarization are consistent with impairment of delayed rectifier potassium channels observed in animal models of iron overload.³³

In our studied population, bradycardia was a relatively rare but specific indicator for cardiac iron. Indeed patients with cardiac iron overload had a significantly lower heart rate than the group with non-detectable cardiac iron, which was in accordance with the results of Detterich et al.²⁹ Although this observation is somewhat surprising given patient anemia, it has previously been described in animal models of iron overload.³³ Atrial iron deposition is reported in thalassemia, but whether the observed changes are due to direct sinus nodal toxicity or increased cardiac vagal tone remains vague.^34,35 Of note, here is our record of a significantly higher incidence of episodes of sinus pauses in cardiac iron overloaded patients, which could point to the possibility of nodal toxicity.

Additionally, high-grade AV block was more commonly encountered in patients with CMRI ≤20 ms. Such a conduction delay may be secondary to compensatory ventricular dilation, increased circulating oxidative radicals like labile plasma iron, or disturbances in electrolytes, vitamins, and trace minerals, which all represent myocardial stress factors.³⁶

We found that in presence of normal LVEF, detection of 10 or more PVCs per hour was the most predictive of preclinical cardiac affection with a PPV of 86.7% and specificity of 96.6%, yet it has a significantly low sensitivity of 30.9, which means that a considerable number of patients still have preclinical cardiac affection and yet they have no detectable PVCs or less than 10 per hour, which was the case in 69% of our patients with preclinical cardiac affection.

Sinus pauses of 2.5 seconds or more; high-grade AV block, and presence of LVDD share the same finding of good PPV with high specificity and low sensitivity. Detection of more than 22 PVCs/24 hours has the best sensitivity and fewer numbers of false-negative cases giving it the best score as a NPV. This implies that recording of 10 or more PVCs, sinus pauses, high-grade AV block and LVDD are highly suggestive of cardiac iron loading, whereas a lower number of PVCs/24 hours holds a higher probability of excluding cardiac iron overload.

We used the positive likelihood ratio to determine the utility of parameters examined. In this regard 24-hour Holter analysis to detect PVCs ≥10/hour and sinus pauses were found to be the most likely parameters to predict preclinical cardiac affection. Most of the published literature is in line with our results stating that number of transfusion per year is significantly correlated to the cardiac affection in β-TM;²⁶ nevertheless, in our work we demonstrated that it has low positive and NPVs. To the best of our knowledge, no available published data discussing the predictive value of the Holter and ECHO parameters in detection of preclinical cardiac affection in β-TM patients, thus marking this work as a start signal for further evaluation of these parameters.

Our results support our hypothesis that monitoring β-TM patients with echo and Holter ECG can help in the detection of preclinical cardiac affection in centers lacking CMRI; however, due to relatively low sensitivity they can not fully replace CMRI. Further work is needed to identify additional simple parameters that can form a diagnostic model with adequate sensitivity.