Non Invasive Evaluation of Hepatic Iron Concentration By Fibroscan in Transfusion-Dependent Egyptian Patients with Chronic Hemolytic Anemia

Abstract

Background: Transient elastography (Fibroscan®) is an ultrasound technique used to measure liver stiffness (LS), and thus assess for liver fibrosis, in patients with various chronic hepatic disorders. It can also be used to predict severity in multiple other diseases that might affect LS such as amyloidosis and possibly conditions associated with iron overload.Objectives: To assess the frequency of liver fibrosis in patients with chronic hemolytic anemia using Transient elastography (Fibroscan®), and to determine the reliability of this tool as a non-invasive method to predict hepatic iron content as compared to liver iron concentration (LIC) measured by magnetic resonance imaging (MRI).Patients and methods: Seventy-five transfusion dependent patients (50 β-thalassemia major;25 sickle cell disease) with a mean age of 13.4±5.2 years in addition to 75 -age and sex matched- healthy children were recruited. All subjects underwent assessment of LS in kilopascals (kPa), by Transient elastography measurement using FibroScan (Echosens, Paris, France І). Steady state serum ferritin (SF), and hepatitis B serologies (HBsAg and antiHB core antibodies) were assessed by enzyme linked immunoassay (ELISA). LIC values, within 6 months' duration, as identified by quantitative MRI of hepatic iron stores as a signal intensity ratio method based on T1 and T2* contrast imaging without gadolinium were retrieved. Informed consent was obtained from patients' legal guardians prior to enrollment in the study.Results: The median SF was 2280 ng/ml (84% had values exceeding 1000 ng/ml). The median LIC was 13.86 mg/g dw (78.7% patients showed LIC above 7 mg/g dw). The median cardiac T2* was 30.8 ms (3 patients had values below 20). Fifty-two (69.3%) patients were categorized as F0-1 and 21 (28%) were stage F2, 2 (1.3%) were stage F3, and 2 patients had severe fibrosis. The mean and median fibroscan (FS) values were 6.19 ±1.76 kPa and 5.9 kPa (range 3 to 14.1) respectively. Patients had significantly higher mean FS compared to control group (p ˂0.001). Patients with no or mild fibrosis (F0-1) had lower FS values (5.3kPa) compared to patients with fibrosis grades 2-4 (p ˂0.001). FS values were not affected by disease type (thalassemia or sickle cell disease), age (above 12 years), or HCV sero-positivity. FS values correlated with SF (r=0.410, p˂ 0.001). Simple regression analysis of the two variables suggested that changes in SF were associated with minimal but significant changes in FS values (p=0.04) with good agreement (kappa =0.324, p=0.003). LIC did not differ in relation to grade of fibrosis (p>0.05), did not correlate with FS values (r= 0.014, p=0.908), and no changes in FS were expected with LIC changes on regression analysis (p=0.466) with low agreement between LIC and FS at cutoff value 5.3 kPa (kappa = 0.015, p=0.9). Sensitivity and specificity of FS values to predict LIC were high at cutoff values ranging between 3.2 to 3.75 kPa but decreased markedly at higher cutoff values. On comparing sensitivity and specificity of FS values in prediction of iron overload at different cutoff values by ROC curve, it could not significantly predict iron overload (p=0.7). No correlations were found between LIC and other variables including SF (r=0.2), and changes in SF were not significantly associated with changes in LIC values (p =0.089). However, sensitivity and specificity of SF in predicting LIC were good at cutoff 1003.85 ng/ml but decreased markedly at higher cutoff values. Comparing its sensitivity and specificity to that of SF in the prediction of iron overload at different cutoff values by ROC curve, FS could not predict iron overload accurately (p=0.9) and the degree of agreement between these two variables as indicators of iron overload was low (kappa=0.063, p=0.478).Conclusion: Fibroscan could be a valuable tool to assess the degree of liver fibrosis in patients with elevated SF, but it does not appear to reliably predict LIC in such group of patients especially with severe iron overload. FS values were not affected by disease type, age above 12 years, or HCV sero-positivity. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.