short report
Noninvasive assessment of liver fibrosis in thalassaemia major patients by transient elastography (TE) – lack of interference by iron deposition
Vito Di Marco,1 Fabrizio Bronte,1 Daniela Cabibi,2 Vincenza Calvaruso,1 Giuseppe Alaimo,1 Zelia Borsellino,3 Francesco Gagliardotto,3 Piero Luigi Almasio,1 Marcello Capra3 and Antonio Craxı`1 1
Dipartimento Biomedico di Medicina Interna e Specialistica (Di.Bi.M.I.S.), Unita` Complessa di Gastroenterologia ed Epatologia, University of Palermo, 2Dipartimento di Diagnostica di
Laboratorio, Servizio di Istologia Patologica, University of Palermo, and 3Unita` Complessa Ematologia-Emoglobinopatie, Ospedale Pediatrico ‘‘Giovanni Di Cristina’’, ARNAS Civico, Palermo, Italy
Summary The correlation between liver stiffness, measured by transient elastography, liver fibrosis, using the histological METAVIR score, and iron overload, measured by atomic absorption spectrometry was evaluated in 56 homozygous-b-thalassaemics. Liver stiffness increased proportionally to liver fibrosis staging (r = 0Æ70; P > 0Æ001) independently of liver iron concentration (r = 0Æ01; P = 0Æ932). The area under the receiver-operating characteristic curve for prediction of cirrhosis was 0Æ997 (95% confidence interval [CI]: 0Æ925–1Æ000) with cut-off of 13 kPa with 100% sensitivity (95% CI: 69Æ0–100Æ0) and 95% specificity (95% CI: 84Æ2–99Æ3). Transient elastography is a reliable non-invasive tool for diagnosing advanced liver fibrosis in homozygous-b-thalassaemics, regardless of the degree of iron overload. Keywords: cirrhosis, liver fibrosis, liver iron concentration, thalassaemia, transient elastography.
Received 8 July 2009; accepted for publication 17 September 2009 Correspondence: Dr Vito Di Marco, Dipartimento Biomedico di Medicina Interna e Specialistica, (Di.Bi.M.I.S.), Cattedra di Gastroenterologia, University of Palermo, Piazza delle Cliniche 2, 90127 Palermo, Italy. E-mail:
[email protected]
Liver biopsy is considered the gold-standard method for evaluating the grade of inflammation, the stage of fibrosis and to measure the liver iron concentration (LIC) of hepatic tissue in patients with homozygous b-thalassaemia (Angelucci et al, 1995, 2000). In the last few years, liver stiffness measurement (LSM) by transient elastography (TE) has been shown to be closely related to the degree of hepatic fibrosis assessed by biopsy in chronic liver diseases; it has been extensively validated in chronic hepatitis C and is a reasonably accurate tool for detection of severe fibrosis and cirrhosis (Castera et al, 2008; Rockey, 2008). The possible influence of liver iron overload on LSM is still unknown. In order to assess whether TE can replace liver biopsy as a tool for assessing liver fibrosis First published online 23 November 2009 doi:10.1111/j.1365-2141.2009.07996.x
in subjects with severe iron overload and frequently infected with hepatitis C virus (HCV), we evaluated this technique prospectively in a cohort of patients with transfusiondependent homozygous b thalassaemia undergoing liver biopsy.
Patients and methods We studied transfusion-dependent thalassaemics prospectively referred to our liver unit for evaluation of their liver pathology. Patients were receiving regular blood transfusions to maintain pre-transfusion haemoglobin values at 90–95 g/l and were treated with iron chelation aimed at keeping ferritin values below 2500 lg/l. Liver biopsy was performed either during
ª 2009 Blackwell Publishing Ltd, British Journal of Haematology, 148, 476–479
Short Report splenectomy for management of thalassaemia, before starting antiviral therapy for chronic HCV hepatitis or before being switched to oral iron chelation drugs. The TE was done at the time of percutaneus biopsy or before splenectomy. The study was performed in accordance with the principles of Good Clinical Practice and was approved by the hospital’s Ethics Committee. Serum aspartate aminotransferase (AST), alanine aminotransferase (ALT) and ferritin were measured at the time of liver biopsy. Anti-HCV-positive patients were tested for qualitative serum HCV-RNA by polymerase chain reaction (Amplicor HCV; Roche Molecular Systems, Basel, Switzerland). Percutaneus liver biopsy was performed using a 16 G Menghini needle. Hepatic inflammation and fibrosis were scored using the METAVIR score system (Bedossa & Poynard, 1996). LIC was measured by atomic absorption spectrometry on fresh tissue cores using the Spectra 880 (Varian, Australia) and results were expressed as mg of iron per gram of liver, dry weight. (Villeneuve et al, 1996). The LSM was measured by TE (FibroScan; Echosens, Paris, France) (Sandrin et al, 2003). Only patients who had at least 10 validated measurements with a success rate of at least 60% and an interquartile range (IQR) < 20% of median values (Castera et al, 2008) were considered reliable (Fraquelli et al, 2007). Data were analysed using the Statistical Package for the Social Sciences (spss) v.13.0 for Windows software (SPSS Inc., Chicago, IL, USA). We used parametric test (t-test) for variables with normal distribution and nonparametric test (Mann–Whitney) for iron overload indices. Chi-square analysis was used for dichotomous or categorical variables. The correlation between staging of liver fibrosis, LIC and LSM was evaluated by Spearman’s rank correlation coefficient. Multiple linear regression model was used to assess the relationships between biochemical, virological, histological variables and TE. Significant variables on univariate analysis (P < 0Æ05) were included in multivariate model. The diagnostic performance of the LSM to detect cirrhosis (METAVIR stages F0–F3 versus stage F4) was assessed by receiver operating characteristic (ROC) curves using the MEDCALC programme. The area under the ROC curve (AUROC), and 95% confidence intervals (CI), were used as indices of accuracy.
Results Fifty-six consecutive patients (45 adults and 11 children) with a suitable liver biopsy entered the study. Forty-five adult patients had a mean age of 32 ± 8Æ4 years (range 19–51), their mean body mass index (BMI) was 24 kg/m2 (range 17Æ5–38Æ5) and 23 (51%) were anti-HCV positive and HCV-RNA positive. All children were anti-HCV and HCV-RNA negative and underwent liver biopsy during splenectomy. Their mean age was 9 ± 1Æ4 years (range 8–12) and their mean body weight was 26Æ9 ± 4Æ1 kg (range 21–35) with a percentile weight of 38Æ7 ± 12Æ4 (range 20–55). In adult patients the median ferritin value was 1573 lg/l, with a median LIC 3Æ5 mg/g liver, dry weight (range 0Æ2–22 mg). In children undergoing splenectomy the median value of
ferritin was 2870 lg/l, and the median value LIC was 8Æ7 mg/g liver, dry weight. TE was not determinable in three patients with a BMI > 31 kg/m2; these patients were excluded from the analysis. The mean length of the liver fragment given for histology was 17 mm (range 11–31). HCV-RNA-positive patients had higher AST and ALT values, and more severe inflammation and fibrosis on biopsy (Table I). The mean number of valid measurements per patient was 11Æ5 (range 10– 18), the success rate was 92Æ3% (range 63–100) and the median IQR was 1Æ1 (range 0Æ2–5Æ8). Mean stiffness value was 7Æ9 ± 6Æ8 kPa (range 2Æ8–24Æ5) and there was a significant difference between patients with chronic hepatitis C and patients free of infection (11Æ4 kPa vs. 5Æ2 kPa; P < 0Æ001). LSM increased proportionally to METAVIR stage, with a highly Table I. Demographic, biochemical and histological features in patients with and without chronic hepatitis C.
Gender (M/F) Age (years) BMI Serum ferritin (lg/l, mean) AST (IU/ml, mean) ALT (IU/ml, mean) GGT (IU/ml, mean) PLT (·109/l, mean) Total bilirubin (lmol/l, mean) HCV genotypes (n, %) 1b 2 a/2c 3a 4c/4d Histological grading A0 A1 A2 A3 Histological staging F1 F2 F3 F4 LIC (mg/g liver, dry weight, median) Steatosis
