Correlation between bone mineral densitometry and liver/heart iron overload evaluated by quantitative T2* MRI

Abstract

Background

Osteomalacia and osteoporosis are a common clinical problem in patients with beta thalassemia major. The purpose of this study is to assess the possible correlation of excess iron in liver and heart of beta thalassemia patient with bone mineral density (BMD).

Materials and methods

We performed a prospective cross-sectional study on 80 patients with beta thalassemia major and inermedia. The patients were stratified into normal BMD and those with abnormal BMD test based on T-score. All patients underwent magnetic resonance imaging examinations of heart and liver. Multi-echo fast gradient-echo sequence was used and T2* values were calculated based on the Royal Brompton Hospital protocol. Correlations between T2* values in heart and liver as well as serum ferritin levels with femoral and lumbar BMD were investigated.

Results

There were no significant differences between patients with and without osteopenia or osteoporosis with respect to age and sex. Patients with abnormal BMD had lower T2* values in the heart. Serum ferritin and heart iron load (based on T2*) were negatively correlated with BMD.

Conclusion

Serum ferritin and heart iron content are good indicators of BMD in patients with beta thalassemia major and intermedia.

Keywords:

• Magnetic resonance imaging
• Iron deposition
• Ferritin
• Beta thalassemia
• T2* value
• Bone mineral density

Introduction

Beta thalassemic patients have a wide variety of musculoskeletal problems due to anatomic proximity of bones and joints to the active centers of hematopoiesis.¹ These include fractures,^1–3 premature epiphyseal fusions,¹ and thalassemic osteoarthropathy.⁴

Most of these patients^1–3 have been previously reported to sustain fractures of the long bones as a result of minor, direct or indirect trauma, especially in the more severe variety of beta thalassemia major.⁵ These problems may have occurred as a result of osteopenia or osteoporosis (OOS), which are common in beta thalassemia.^6–8 Beta thalassemia is usually considered as one of the endocrinopathies, which is caused by hypogonadism. However some studies have shown the role of iron overload in the pathogenesis of OOS.⁹ Here we aimed to study the correlation of iron overload in the liver and heart of the thalassemic patients with bone marrow density (BMD). We used magnetic resonance imaging (MRI) in order to define the iron content of different organs.

Materials and methods

Participants

We performed a prospective cross-sectional study on 80 beta thalassemia cases (major and intermedia). The study was conducted from January 2010 to July 2011, in Noor Medical Imaging Center, Tehran, Iran. Patients with documented beta thalassemia major were consecutively enrolled from the Zafar hematology clinic, affiliated with our university. Beta thalassemia major was diagnosed according to hemoglobin electrophoresis. All the patients were on regular transfusion therapy, none had major clinical problem. Participants with infections, metabolic diseases, or women on hormone replacement therapy were excluded. A single physician examined all participants for signs of growth retardation, movement disorder, and irregular menstruation. The study was approved by the local ethics review committee of Noor Medical Imaging Center. After explaining the study process in detail to the patients and controls, written informed consent was obtained from all participants. Serum ferritin also accessed by use of enhanced chemiluminescence method (expressed as mg/g/dw).

MRI technique

All patients underwent MRI examinations on a 1.5 Tesla Symphony scanner (Siemens, Erlangen, Germany). A standard RF body coil for liver and heart were used. A multi-echo fast gradient-echo sequence was used and T2* values were measured based on the Royal Brompton Hospital protocol.¹⁰ The MRI T2* of the liver was determined using a single 10-mm slice through the center of the liver scanned at 12 different echo times (TE) (1.3–23 ms). Each image was acquired during a 11–13 seconds breath-hold using a gradient-echo sequence (repetition time 200 ms, flip angle 20°, base resolution matrix 128 pixels, field of view 39.7 × 19.7 cm², sampling bandwidth of 125 kHz).

For the measurement of myocardial T2*, a single short axis mid-ventricular slice (10 mm) was acquired at eight separate echo times (3–22·6 ms). The repetition time between radio frequency (RF) pulses was between 125 and 223 ms, depending on the echo time used. A gradient-echo sequence was used (flip angle 20, base resolution matrix 256 pixels, field of view 29.8–39.8 cm, sampling bandwidth of 810 Hz/pix). The repetition time was adjusted to the patient's heart rate. Each image was acquired during a 12–17 seconds of breath-hold. A gating delay time of 0 ms after the R-wave of electrocardiogram and acquisition windows were chosen in order to obtain myocardial images in a consistent position in the cardiac cycle irrespective of the heart rate. A full-thickness region interest was measured in the left ventricular myocardium, encompassing both epicardial and endocardial regions. This was located in the septum, distant from the cardiac veins, which could cause susceptibility artifacts. All T2* quantization performed by using dedicated homemade software (Noor Medical Imaging Centre, Tehran, Iran). Liver iron was extracted from the exponential equation: Fe liver = EXP(2.66409 − (1.06611*LN(T2*))).

Bone densitometry

Bone densitometry of lumbar spine (L2–L4) and right femoral neck were measured by Quantitive CT (QCT) with Sensation 64 Slice scanner (Siemens) and analyzed by PRO QCT 3D Volumetric QCT bone densitometery of the hip and spine Software (Mindways, Austin, USA). Bone densitometry was expressed as gram per square centimeter and T-score and Z-score. OOS were delineated according to World Health Organization criteria. The maximum interval between assessment of iron overload and serum ferritin and performance of bone densitometry was three months.

Statistical analysis

Statistical analysis was performed by SPSS software (version 16.0, SPSS Inc., Chicago, IL, USA). Continuous variable are expressed as mean ± standard error of mean. Comparison of the variables between patients and controls was made using χ² analysis for categorical variables and students’ sample t-test for normally distributed continuous variables. P < 0.05 was considered statistically significant. The patients were stratified into normal BMD (T-score >− 1) and those with OOS (T-score ≤− 1). For further analysis we divided the patients into under and above 20 years of age.

Results

There were 80 patients, with the mean age of 26 years old, while 36 of them were women.

Forty-nine of them had thalassemia major and 31 had thalassemia intermedia.

Twelve of the patients with thalassemia intermedia and 16 of the thalassemia major patients had decreased femoral BMD. There was no significant difference in age, sex, serum ferritin, liver/heart iron, and BMD between patients with thalassemia major and intermedia. Characteristics of patients with thalassemia major and thalassemia intermedia are presented in Table 1.

Table 1. Characteristics of patients with thalassemia major and thalassemia intermedia

	Thalassemia major (N = 49)	Thalassemia intermedia (N = 29)	P value
Age of patient	24.39 ± 6.03	29.21 ± 7.35	<0.05
BMD lumbar	166.01 ± 36.39	146.85 ± 44.90	<0.05
BMD femoral	0.81 ± 0.16	0.77 ± 0.17	NS
Ferritin	1875.63 ± 1722.24	1392.39 ± 1437.60	NS
T2* heart	24.88 ± 13.62	29.68 ± 12.8	NS
Fe liver	5.36 ± 3.39	4.23 ± 2.33	NS

We then stratified the patients into those with normal BMD and those with OOS. There was no significant difference in age, sex, liver iron, and heart iron of the patients between those of normal BMD and those with OOS (Table 2).

Table 2. Characteristics of patients stratified according to the normal BMD and those with OOS

	Normal BMD (n = 50)	Osteomalcia/osteoporosis (n = 30)	P value
Age of patient	24.55 ± 0.75	28.69 ± 1.56	<0.001
Female (n, %)	22 (43%)	14 (48%)	NS
Thalassemia major (n, %)	33 (64%)	16 (55%)	NS
T-score lumbar	0.30 ± 0.142	−2.00 ± 0.19	<0.001
BMD lumbar	180.56 ± 3.76	120.06 ± 5.35	<0.001
BMD femoral	0.88 ± 0.015	0.65 ± 0.026	<0.001
Ferritin	1912.21 ± 247.95	1293.39 ± 249.76	NS
T2* heart	27.31 ± 1.73	26.02 ± 2.78	NS
T2* liver	5.54 ± 1.02	3.43 ± 0.408	NS
Fe liver	5.07 ± 0.44	4.55 ± 0.55	NS

We then studied the correlation of the studied variables within patients with normal BMD and those with OOS (Table 3). Serum ferritin, was negatively correlated with femoral BMD and lumbar BMD in patients with OOS (r = −0.56; P < 0.001) and (r = −0.055; P < 0.001), respectively, while there was no such a correlation in those with normal BMD (Table 3).

Table 3. Correlation of BMD with liver, heart, and serum iron content, in patients with OOS and normal BMD

	Patients with OOS			Patients with normal BMD
	BMD femoral	T2* heart	T2* liver	BMD femoral	T2* heart	T2* liver
Ferritin	−0.52*	−0.44**	−0.32	0.12	−0.46*	−0.36*
T2* heart	0.42**	1	0.24	0.10	1	0.13
T2* liver	0.01	0.24	1	0.02	0.13	1
Age of patient	−0.008	0.25	0.12	−0.05	0.28**	−0.04
BMD lumbar	0.55*	0.33	0.13	0.46*	−0.03	−0.11

*Correlation is significant at the 0.05 level (two-tailed).

**Correlation is significant at the 0.01 level (two-tailed).

We then studied these correlation while stratifying the patients into below (Table 4) and above 20 (Table 5) years of age.

Table 4. Correlation of BMD with liver, heart, and serum iron content, in patients with OOS and normal BMD (age <20)

	Patients with OOS				Patients with normal BMD
	BMD femoral	T2* heart	T2* liver	BMD lumbar	BMD femoral	T2* heart	T2* liver	BMD lumbar
Ferritin	0.231	−0.085	−0.584	−0.260	−0.703	−0.474	−0.368	−0.603
T2* heart	−0.260	1	0.365	0.165	0.057	1	0.112	0.581
T2* liver	−0.413	0.365	1	0.523	−0.017	0.112	1	−0.497
Age of patient	0.441	0.410	0.217	0.379	−0.641	0.556	−0.278	−0.390
BMD lumbar	−0.493	0.165	0.523	1	0.064	0.581	−0.497	1
Fe liver	0.744	−0.538	−0.811	−0.580	−0.076	−0.414	−0.772*	0.110

Table 5. Correlation of BMD with liver, heart, and serum iron content, in patients with OOS and normal BMD (age ≥20)

	Patients with OOS				Patients with normal BMD
	BMD femoral	T2* heart	T2* liver	BMD lumbar	BMD femoral	T2* heart	T2* liver	BMD lumbar
Ferritin	−0.561*	−0.522**	−0.404	−0.55*	0.239	−0.436**	−0.310	0.466*
T2* heart	0.485**	1	0.493**	0.447**	0.130	1	0.156	−0.172
T2* liver	0.362	0.493**	1	0.226	0.002	0.156	1	−0.175
Age of patient	−0.002	−0.092	−0.317	−0.371	−0.220	0.182	0.089	−0.292
BMD lumbar	0.580*	0.447**	0.226	1	0.470*	−0.172	−0.175	1
Fe liver	0.353	−0.495**	−0.710*	−0.240	−0.041	−0.460*	−0.553*	0.370**

*Correlation is significant at the 0.05 level (two-tailed).

**Correlation is significant at the 0.01 level (two-tailed).

Discussion

Our findings clearly demonstrated that there was a negative correlation of serum ferritin and heart iron with BMD in patients with OOS. We also indicate that low BMD is often present in patients with thalassemia and this is significantly correlated with the amount of iron overload in these patients. Although so many other studies haves shown the correlation of BMD with ferritin and liver iron, none have mentioned this correlation with heart iron. This is partly due to the lack of diagnostic facilities to measure cardiac iron content. We used T2* value to study the amount of heart iron. To date we are unaware of any study demonstrating the correlation of BMD with iron overload in patients with thalassemia, using QCT and T2* values.

The underlying pathogenic mechanisms contributing to the development of bone disease in thalassemia are multiple and complicated.¹¹ Delayed bone maturation and focal osteomalacia are the main causes of bone disease in suboptimally blood-transfused thalassemics with iron overload.^9,12 Recent studies suggest that the reduced osteoblastic activity, which is believed to be the basic mechanism of bone loss in thalassemia, is accompanied by a comparable or even greater increase in bone resorption.¹³ The metal suppresses osteoblast formation of bone and may also stimulate osteoclast resorption of bone. Iron also inhibits anterior pituitary synthesis of gonadotrophs. We have previously shown the iron overload syndrome in the brain of the thalassemic patients.¹⁴ It is suggested that brain iron results in depressed formation of gonadal hormones. The tendency of iron-loaded persons to become osteoporotic may be enhanced by gonadal hormone deficiency.^15,16 Consistently hormone replacement therapy have shown promising outcome in prevention of osteoporosis.¹⁷

This study showed that in osteoporotic patients, femoral BMD is negatively correlated with ferritin and heart iron content. Elevated liver iron concentration in patients with beta thalassemia intermedia is a marker of increased vascular, endocrine, and bone disease.¹⁸ These findings may imply the role of iron content on vascular function which may result into the osteoporosis.

High prevalence of osteoporosis in our thalassemic patients is in gross agreement with those reported in the earlier literature. Similar results from other studies have also been reported.^19,20 Jensen et al. studied 82 patients who were well transfused and who received regular desferrioxamine. In their study, the overall prevalence of patients with ‘severely low’ bone mass was 51% and those with ‘low’ bone mass was 45%.²¹ This is in line with the current finding, however we showed that this may have had resulted from the iron overload. It is suggested that regular iron monitoring, could be an important and significant marker for the diagnosis of osteoporosis in patients with thalassemia, as it is simple, has a lower cost, and is easier to measure.

In conclusion, from a different perspective our results corroborate the role of endocrinopathies in thalassemic osteopathy, and once again underline the crucial importance of an early and multi-disciplinary intervention in preventing bone complications in thalassemic patients. The principal limitation of the present study is its cross-sectional nature that precludes the determination of the direction of causality; however, we took advantage of a relatively large sample size and close similarity between groups in most of the potentially cofounding variables.