GASTROENTEROLOGY 2011;141:1202–1211
CLINICAL—LIVER
Improvement in Liver Pathology of Patients With ␤-Thalassemia Treated
With Deferasirox for at Least 3 Years
YVES DEUGNIER,* BRUNO TURLIN,‡ MARTINE ROPERT,§ M. DOMENICA CAPPELLINI,㛳 JOHN B. PORTER,¶
VANESSA GIANNONE,# YIYUN ZHANG,# LOUIS GRIFFEL,# and PIERRE BRISSOT*
*Liver Disease Unit and Inserm U-991, University Hospital Pontchaillou, Rennes, France; ‡Department of Pathology, Biological Resource Center and Inserm U-991,
University Hospital Pontchaillou, Rennes, France; §Laboratory of Biochemistry, University Hospital Pontchaillou, Rennes, France; 㛳Universitá di Milano, Ca Granda
Foundation IRCCS, Milan, Italy; ¶University College London, London, UK; #Novartis Pharmaceuticals Corporation, East Hanover, New Jersey
CLINICAL LIVER
See editorial on page 1142.
BACKGROUND & AIMS: Most data on the effects of
iron chelation therapy for patients with liver fibrosis come
from small studies. We studied the effects of the oral iron
chelator deferasirox on liver fibrosis and necroinflamma-
tion in a large population of patients with iron overload
␤-thalassemia. METHODS: We studied data from 219
patients with ␤-thalassemia, collected from histologic
analyses of biopsy samples taken at baseline and after at
least 3 years of treatment with deferasirox. Treatment
response was assessed from liver iron concentrations at
baseline and the end of the study. Liver fibrosis, necroin-
flammation, and markers of iron overload and liver en-
zymes were recorded. Patients were also assessed, by sero-
logic analysis at baseline, for hepatitis C virus infection.
RESULTS: By the end of the study, stability of Ishak
fibrosis staging scores (change of ⫺1, 0, or ⫹1) or im-
provements (change of ⱕ ⫺2) were observed in 82.6% of
patients; Ishak necroinflammatory scores improved by a
mean value of ⫺1.3 (P ⬍ .001). Improvements in fibrosis
stage and necroinflammation were independent of hepa-
titis C virus exposure or reduction in liver iron concen-
tration defined by the response criteria. Absolute changes
in concentrations of liver iron by the end of the study did
not correlate with improved Ishak fibrosis or necroinflam-
matory scores. CONCLUSIONS: Deferasirox treatment
for 3 or more years reversed or stabilized liver fibrosis
in 83% of patients with iron-overloaded ␤-thalasse-
mia. This therapeutic effect was independent of re-
duced concentration of liver iron (defined by the re-
sponse criteria) or previous exposure to hepatitis C
virus.
Keywords: Liver Disease; Clinical Trial; Cirrhosis; Therapy.
T he liver is the principal site of iron accumulation in
patients with transfusional iron overload. Storage of
excess iron in the liver can lead to tissue damage, collagen
formation, and portal fibrosis within 2 years of the start
of transfusion therapy, eventually resulting in cirrhosis
and ultimately multiple organ damage.1,2 In addition to
the exacerbating effects of excess iron on liver fibrosis,
iron overload together with hepatitis C virus (HCV) infec-
tion has a synergistic effect on hepatic fibrogenesis.3,4
Despite this serious clinical sequela of iron overload, trials
evaluating liver fibrosis during iron chelation therapy and
the role of HCV are limited to a number of small stud-
ies.5–7 Indeed, evidence supporting the ability of any ther-
apeutic agent to reverse liver fibrosis is limited.8 –13 In
order to address and further explore these questions in a
larger patient population, a subanalysis of data obtained
from 2 large studies of therapy with the iron chelator
deferasirox has been performed (see Appendix). Study 107
was a 1-year, randomized, controlled trial of deferasirox vs
deferoxamine (DFO) in patients aged at least 2 years with
␤-thalassemia major.14 Study 108 was a 1-year noncom-
parative trial of deferasirox in patients with various trans-
fusion-dependent anemias, including ␤-thalassemia.15 Ex-
tension phases of these studies (107E and 108E)
continued for an additional 4 years. Long-term efficacy
and safety data for up to 5 years have been reported from
studies 107E and 108E. Patients experienced substantial
reductions in iron burden, with an increasing proportion
achieving therapeutic goals of liver iron concentration
(LIC) ⬍7 mg Fe/g dry weight (dw) and serum ferritin
levels ⱕ1000 ng/mL over time.16 In pediatric patients,
deferasirox treatment for up to 5 years was effective in
decreasing serum ferritin to levels of approximately 1000
ng/mL without adverse effects on growth and develop-
ment.17 During these studies, liver biopsy samples were
collected, which facilitated the evaluation of liver pathol-
ogy during deferasirox treatment.
Data presented in this subanalysis are for at least 3
years of treatment with deferasirox, as previous studies
have demonstrated that a shorter time period of 1 year
Abbreviations used in this paper: ALT, alanine aminotransferase;
DFO, deferoxamine; dw, dry weight; EOS, end of study; HCV, hepatitis C
virus; HIS, hepatocytic iron score; LIC, liver iron concentration; LPI,
labile plasma iron; PIS, portal iron score; SIS, sinusoidal iron score; TIS,
total iron score.
© 2011 by the AGA Institute
0016-5085/$36.00
doi:10.1053/j.gastro.2011.06.065
October 2011
Table 1. Response Criteria Based on LIC From Biopsy
Baseline LIC Success, if LIC at EOS
(mg Fe/g dw) (Group A)
⬍7 1 to ⬍7 mg Fe/g dw and increase ⬍1 mg Fe/g dw
ⱖ7 to ⬍10 1 to ⬍7 mg Fe/g dw
ⱖ10 Decreases in LIC ⱖ3 mg Fe/g dw
aFailuregroup had lower baseline LIC (10.7 vs 18.3 mg Fe/g dw) and therefore received lower doses of deferasirox, which were often insufficient
to achieve an overall reduction in LIC.
might not have been sufficient to see a modification of
fibrosis staging during treatment with either deferasirox
or DFO.18 An assessment of fibrosis progression has been
made in patients according to LIC response success and
previous HCV exposure. To our knowledge, this is the first
long-term analysis to assess the efficacy of an iron chela-
tor on liver fibrosis in a large cohort of iron-overloaded
patients with ␤-thalassemia.
Patients and Methods
Study Design and Patient Population
Study design and protocols for studies 107 and 108 have
been described previously.14,15 In study 107, patients were ran-
domized to receive deferasirox or DFO (ratio 1:1) for 1 year,
followed by deferasirox for an additional 4 years. Patients who
received DFO during the first year were referred to as “crossover”
cohort. Baseline in this analysis refers to start of deferasirox
treatment, which is the crossover baseline for the crossover
cohort. In study 108, patients received deferasirox only, both in
the 1-year core and 4-year extension phases of this study. In both
studies, enrolled patients were aged at least 2 years of age, of
either sex, received 8 or more blood transfusions per year, and
with transfusional hemosiderosis as indicated by LIC values of at
least 2 mg Fe/g dw. For either trial, patients were excluded if they
had one of the following conditions: an alanine aminotransfer-
ase (ALT) level ⬎250 IU/L in the year before enrollment, serum
creatinine above the upper limit of normal at screening, sero-
logical evidence of chronic hepatitis B, a history of human
immunodeficiency virus seropositivity, or clinical evidence of
active HCV infection (liver pathology, positive HCV total anti-
body test, and abnormal serum aminotransferase levels). Further
exclusion criteria were as described previously.14,15 In the current
subanalysis, patients with ␤–thalassemia who had treatment
with deferasirox for at least 3 years and had liver biopsy assess-
ment (defined as having LIC, Ishak grading, or Ishak staging
assessment) after at least 3 years of treatment were included.
Studies 107 and 108 conformed to Good Clinical Practice and
received institutional review board or ethics committee approval.
Written informed consent was obtained from all patients (or
parents/guardians) before participation in any study proce-
dures.14,15
Dosing
The dosing algorithm used to determine initial defera-
sirox and DFO dosages (deferasirox 5⫺30 mg/kg/d; DFO 20 to
ⱖ50 mg/kg, 5 days per week) according to the baseline LIC of
each patient has been described previously.14,15 During the ex-
tension trials, dosage increases or decreases of 5–10 mg/kg/d
could be made every 3 months based on trends in serum ferritin
levels and safety markers. Dosage adjustments could also be
IMPROVEMENT IN LIVER PATHOLOGY WITH DEFERASIROX 1203
Failure,a if LIC at EOS
(Group B)
⬍1 mg Fe/g dw or ⱖ7 mg Fe/g dw or increase ⱖ1 mg Fe/g dw
⬍1 mg Fe/g dw or ⱖ7 mg Fe/g dw
Decreases in LIC ⬍3 mg Fe/g dw
based on transfusion requirements at the investigator’s discre-
tion.
Assessments
CLINICAL LIVER
LIC was determined at the start of deferasirox treatment
and at end of study (EOS; last available result during the study)
by liver biopsy using a standardized chemical methodology19
with detection by atomic absorption spectrometry.20,21 Biopsies
were fixed and paraffin-embedded, and iron content was re-
ported as dw on de-waxed samples. Biopsy specimens with a
total dw ⬍0.4 mg were not evaluated for iron content.14 Treat-
ment response success was defined according to baseline (start
of deferasirox dosing) and EOS LIC measurements defined ac-
cording to the core 107 study (Table 1).14 LIC values ⬎7 mg Fe/g
dw have been reported to be associated with increased morbidity
and mortality;22 hence maintenance or reduction of LIC below
this end point was deemed a desirable response (LIC response
success, Group A). Reduction in LIC to ⬍1 mg Fe/g dw was
undesirable, potentially exposing patients to overchelation (LIC
response failure, Group B). LIC values are expressed as mg Fe/g
dw. For conversion to SI units (␮mol/g), the figure presented
can be multiplied by a conversion factor of 17.9.
Serum ferritin and ALT were assessed using standard blood
chemistry examinations. Total iron score (TIS) was used to
semi-quantitatively assess iron load in the liver and was mea-
sured at baseline and EOS using 3-␮m tissue sections stained
with Prussian blue dye.23 TIS was derived from the iron load
observed in hepatocytes (hepatocytic iron score [HIS] range,
0⫺12), sinusoidal cells (sinusoidal iron score [SIS] range, 0⫺4)
and main structures of the portal tracts (portal iron score [PIS]).
A heterogeneity factor (1, 2, or 3) was then applied, based on the
overall appearance of the tissue, to provide TIS, calculated as
(HIS ⫹ SIS ⫹ PIS) ⫻ (heterogeneity factor/3) (range, 0⫺60). The
liver iron ratio, used to assess the relative amount of hepatocytic
to total liver iron, was calculated as HIS/(HIS ⫹ SIS ⫹ PIS) ⫻
100%. Fibrosis staging was performed according to Ishak scale
from 0 (no fibrosis) to 6 (cirrhosis, probable or definite).24 Liver
inflammation and necrosis/damage was assessed according to
the Ishak necroinflammatory grading system with an overall
scoring range from 0 (no inflammation) to 18 (severe).24 HCV
status was assessed at baseline by the presence or absence of
HCV antibodies (evidence of HCV exposure); HCV RNA was
unavailable. HCV–positive patients were defined as being HCV
positive on or before the start of deferasirox treatment. The
primary objective of this subanalysis was an assessment of the
efficacy of deferasirox on liver fibrosis and necroinflammation in
patients with ␤-thalassemia.
Statistical Methods
The percentage of LIC response for each treatment
was calculated as described in the core 107 study.14 Unless
 1204 DEUGNIER ET AL GASTROENTEROLOGY Vol. 141, No. 4

Table 2. Demographics and Baseline Characteristics of Patients Who Had Histological Biopsy Data at Baseline and at the
End of at Least 3 Years Treatment With Deferasirox by Study Cohort, and Baseline Efficacy Measurements for
Patients With Baseline and EOS LIC Assessments by LIC Response Criteria Category
Study 107 Study 108

Deferasirox Crossovera Deferasirox All patients

Characteristic (n ⫽ 106) (n ⫽ 94) (n ⫽ 19) (n ⫽ 219)
Mean age, y, (range) 15.0 (2–36) 15.0 (3–43) 22.1 (4–49) 15.6 (2–49)
Male/female 55/51 56/38 8/11 119/100
Race (Caucasian/Asian/other), n 100/1/5 88/2/4 9/4/6 197/7/15
History of hepatitis at start of deferasirox
treatment, n (%)
No hepatitis B or C 81 (76.4) 77 (81.9) 18 (94.7) 176 (80.4)
Hepatitis B 10 (9.4) 7 (7.4) 0 (0.0) 17 (7.8)
Hepatitis C 19 (17.9) 12 (12.8) 1 (5.3) 32 (14.6)
Hepatitis B and C 4 (3.8) 2 (2.1) 0 (0.0) 6 (2.7)
Hepatitis NOS 1 (0.9) 2 (2.1) 1 (5.3) 4 (1.8)
CLINICAL LIVER

Any type of hepatitis 25 (23.6) 17 (18.1) 2 (10.5) 44 (20.1)

Baseline liver parameters (mean ⫾ SD)
LIC, mg Fe/g dw 17.8 ⫾ 10.6 12.6 ⫾ 8.2 19.1 ⫾ 10.3 15.7 ⫾ 9.9
TIS 27.2 ⫾ 11.0 22.0 ⫾ 10.4 28.6 ⫾ 10.3 25.1 ⫾ 11.0
(n ⫽ 104) (n ⫽ 92) (n ⫽ 19) (n ⫽ 215)
Liver iron ratio,b % 66.0 ⫾ 13.9 69.3 ⫾ 15.1 56.0 ⫾ 14.8 66.5 ⫾ 14.8
(n ⫽ 104) (n ⫽ 92) (n ⫽ 19) (n ⫽ 215)
Median baseline serum ferritin, ng/mL (range) 2148 (367⫺11,453) 1716 (273–8529) 4056 (1402–11,698) 2069 (273⫺11,698)
Baseline efficacy measurements for patients with baseline and EOS LIC assessments
Group A (LIC response success) Group B (LIC response failure) Total
(n ⫽ 134) (n ⫽ 76) (n ⫽ 210)
Mean baseline necroinflammatory score (⫾SD) 2.2 ⫾ 1.7 1.6 ⫾ 1.5 2.0 ⫾ 1.6
Mean baseline ALT, IU/mL (⫾SD) 46.2 ⫾ 41.5 30.6 ⫾ 29.0 40.5 ⫾ 38.2
Baseline liver parameters (mean ⫾ SD)
LIC, mg Fe/g dw 18.3 ⫾ 10.7 10.7 ⫾ 5.9 15.5 ⫾ 9.9
Total iron score 27.9 ⫾ 10.9 (n ⫽ 131) 19.5 ⫾ 8.9 (n ⫽ 75) 24.9 ⫾ 11.0 (n ⫽ 206)
Liver iron ratio, % 65.8 ⫾ 12.7 (n ⫽ 131) 68.0 ⫾ 18.7 (n ⫽ 75) 66.6 ⫾ 15.1 (n ⫽ 206)
Median baseline serum ferritin, ng/mL (range) 2379 (536⫺11,453) 1587 (273⫺11,698) 2049 (273⫺11,698)

NOS, not otherwise specified; SD, standard deviation.

aCrossover, patients treated with DFO during the core phase and deferasirox during the extension.
bLiver iron ratio is defined as HIS / (HIS ⫹ SIS ⫹ PIS) ⫻ 100%.

otherwise specified, reported P values were based on 2-sided
significance tests (1-sample Student t test) with an ␣ level
of .05 to test whether the mean change was 0. For serum
ferritin assessments, a sign test P value was used to test
whether the median change was 0, also with a 2-sided ␣ level
of .05. Correlation was assessed using Pearson correlation
coefficients. Efficacy data for the overall population are pre-
sented for those patients with histological biopsy data at
baseline and at the end of at least 3 years of deferasirox.
Efficacy data for patients with LIC response are based on
those patients who had evaluable baseline and EOS LIC
assessments.
for all patients, as some chose not to undergo a liver
biopsy at EOS. From these 219 patients, 210 had evalu-
able baseline and EOS LIC assessments and were clas-
sified as having LIC response success (Group A; n ⫽
134; HCV-positive n ⫽ 20, HCV-negative n ⫽ 114) or LIC
response failure (Group B; n ⫽ 76; HCV-positive n ⫽ 10,
HCV-negative n ⫽ 66). The remaining 9 patients did not
have evaluable LIC for a number of reasons, including
patients with liver biopsies ⬍0.4 mg. Baseline efficacy
measurements for Group A and Group B are shown in
Table 2.

Results Deferasirox Exposure

Patient Characteristics
Of the 671 patients with ␤-thalassemia enrolled
in the core studies, 470 received deferasirox for at least
3 years. Of these patients, 219 had histological biopsy
data at baseline and at the end of at least 3 years of
treatment with deferasirox and, as a result, were eligible
for analysis (Table 2). Paired biopsies were not available
For all patients in this subanalysis (n ⫽ 219), median
duration of deferasirox exposure was 5.1 years (range,
3.8⫺5.6 years). Median deferasirox exposure was comparable
between Group A (5.1 years; range, 3.8⫺5.6 years) and
Group B (4.3 years; range, 4.0⫺5.6 years); and between HCV-
positive (5.1 years; range, 3.9⫺5.5 years) and HCV-negative
(5.1 years; range, 3.8⫺5.6 years) patients.
October 2011
Figure 1. (A) Distribution of patients based on Ishak stages at baseline
and EOS; (B) proportion of patients with Ishak stage improvement, sta-
bility, or worsening by EOS; and (C) scatter plot of absolute changes
from treatment initiation for LIC and Ishak fibrosis staging scores. Only
patients with LIC and Ishak fibrosis staging at both baseline and EOS are
included.
Effects of Deferasirox on Liver Fibrosis
Ishak fibrosis staging scores at baseline ranged
from 0 to 6. The distribution of patients based on Ishak
fibrosis staging scores at baseline and at EOS is shown
in Figure 1A. Comparable distributions of Ishak fibro-
sis staging scores at baseline were observed in Groups A
and B (data not shown). Of all patients in this suba-
nalysis, 55.7% (n ⫽ 122) experienced stability (change
of ⫺1, 0, or ⫹1) with 26.9% (n ⫽ 59) experiencing
improvement (decrease of 2 or more) in their Ishak
fibrosis staging (Figure 1B). Similar improvements were
observed in Group A (59.7% with stability [n ⫽ 80];
26.1% with improvement [n ⫽ 35]) and Group B (48.7%
with stability [n ⫽ 37]; 30.3% with improvement [n ⫽
23]). Overall, 68.0% (n ⫽ 149) experienced a change of
⫺2, ⫺1, or 0. Stability or improvement in fibrosis
staging was observed in 46.7% (n ⫽ 14) and 30.0% (n ⫽
9) of HCV-positive patients and in 57.2% (n ⫽ 103) and
27.2% (n ⫽ 49) of HCV-negative patients, respectively.
As crossover patients had all received 1 year of DFO
therapy on study before initiating deferasirox, we also
analyzed these patients separately from those who had
received deferasirox (deferasirox cohort) from the out-
IMPROVEMENT IN LIVER PATHOLOGY WITH DEFERASIROX 1205
set. Although there was no difference in the proportion
of patients who improved based on LIC response, those
in the deferasirox cohort were more likely to have a
worsening in their fibrosis stage if they did not meet the
LIC response criteria (7% of patients in Group A vs 24%
in Group B) than those who were in the crossover
cohort (4% of patients in Group A vs 8% in Group B;
Supplementary Table 1). In order to evaluate more
rigorously if LIC response to deferasirox correlated with
improvements in Ishak fibrosis staging, a regression
analysis was performed. As shown in Figure 1C, abso-
lute changes in Ishak fibrosis staging did not correlate
with changes in LIC (n ⫽ 195; R ⫽ 0.18). In addition,
absolute changes in LIC did not correlate with changes
in Ishak fibrosis staging in Group A (n ⫽ 123; R ⫽
CLINICAL LIVER
0.17), Group B (n ⫽ 72; R ⫽ 0.17), in HCV-positive
(n ⫽ 26; R ⫽ 0.18) or HCV-negative (n ⫽ 169; R ⫽ 0.25)
patients (data not shown). Improvement in Ishak fibro-
sis staging by 2 or more stages was observed in 26.9%
9(n ⫽ 59) of patients. Similar results were observed in
Group A (26.1%, n ⫽ 35) and Group B (30.3%, n ⫽ 23)
and were regardless of HCV status (HCV-positive:
30.0%, n ⫽ 9; HCV-negative: 27.2%, n ⫽ 49). Worsening
of Ishak fibrosis staging by 2 or more stages was expe-
rienced by 10% (n ⫽ 22) of patients overall; 6% (n ⫽ 8)
of patients in Group A; and 15.8% (n ⫽ 12) of patients
in Group B. Comparisons between Group A and B in
this case should, however, be interpreted with caution,
given the relatively small sample size.
Effects of Deferasirox on Liver
Necroinflammation
For all patients in this subanalysis (n ⫽ 219), Ishak
necroinflammatory grading scores at baseline ranged
from 0 to 8 with a mean of 2.0 ⫾ 1.6 (2.2 ⫾ 1.7 in Group
A; 1.6 ⫾ 1.5 in Group B). By EOS, Ishak necroinflamma-
tory scores improved by a mean of ⫺1.3 (95% confidence
interval: ⫺1.5 to ⫺1.0; P ⬍ .001) in all patients, with a
mean absolute change of ⫺1.5 (95% confidence interval:
⫺1.8 to ⫺1.1; P ⬍ .001) in Group A and ⫺0.8 (95%
confidence interval: ⫺1.2 to ⫺0.4; P ⬍ .001] in Group B
(Figure 2A). Mean Ishak necroinflammatory grading score
at treatment initiation was higher in HCV-positive pa-
tients (3.2 ⫾ 1.9) compared with HCV-negative patients
(1.8 ⫾ 1.5; data not shown). Nonetheless, Ishak necroin-
flammatory grading improved in both HCV-positive
(mean absolute change ⫺1.9 ⫾ 2.2; P ⬍ .001) and HCV-
negative patients (mean absolute change ⫺1.1 ⫾ 1.7; P ⬍
.001; Figure 2A). Regression analysis was also performed
to assess any correlation between change in LIC and
change in Ishak necroinflammatory grading. As shown in
Figure 2B, absolute change in Ishak necroinflammatory
grading did not correlate with change in LIC (n ⫽ 195;
R ⫽ 0.25; Figure 2B). In addition, absolute changes in LIC
did not correlate with changes in Ishak necroinflammatory
grading in either Group A (n ⫽ 124; R ⫽ 0.24); Group B
(n ⫽ 71; R ⫽ 0.04); in HCV-positive (n ⫽ 26; R ⫽ 0.16);
 1206 DEUGNIER ET AL GASTROENTEROLOGY Vol. 141, No. 4
CLINICAL LIVER

Figure 2. (A) Mean absolute change in Ishak necroinflammatory grading by EOS; and (B) scatter plot of absolute changes from treatment initiation
for LIC and Ishak necroinflammatory grading. Only patients with LIC and Ishak necroinflammatory grading at both baseline and EOS are included.
ⴱP ⬍ .001 at EOS compared with baseline. †P ⫽ .002 at EOS compared with baseline.

or HCV-negative (n ⫽ 169; R ⫽ 0.34) patients (data not
shown).
Effects of Deferasirox on Serum ALT
For all patients in this subanalysis (n ⫽ 219), mean
serum ALT (an indicator of hepatocellular damage) levels
decreased from 40.9 ⫾ 37.8 IU/L to 29.6 ⫾ 31.2 IU/L
(mean absolute change ⫺11.3 ⫾ 34.6 IU/L; P ⬍ .001;
mean relative change of ⫺7.4%; P ⫽ .148; Figure 3A).
Mean serum ALT levels at baseline were higher in Group
A (46.2 ⫾ 41.5 IU/L) compared with Group B (30.6 ⫾ 29.0
IU/L); but comparable between HCV-positive (42.4 ⫾ 33.0
IU/L) and HCV-negative (40.2 ⫾ 39.0 IU/L) patients.
Changes in serum ALT mirrored changes in LIC (see
Figure 4) with statistically significant decreases in serum
ALT observed only in Group A (mean absolute change
⫺21.5 ⫾ 35.0 IU/L; P ⬍ .001; Figure 3A). Significant
increases of 6.8 ⫾ 25.2 IU/L (P ⫽ .022) were observed in
Group B. Serum ALT levels improved in both HCV-posi-
tive and HCV-negative patients, but this reduction was
statistically significant only in the latter group (mean
absolute changes of ⫺2.5 ⫾ 25.0 IU/L [P ⫽ .593] and
⫺12.7 ⫾ 35.7 IU/L [P ⬍ .001], respectively; Figure 3A).
Regression analysis revealed a correlation between ab-
solute changes in LIC vs absolute changes in serum ALT
(n ⫽ 210; R ⫽ 0.49; Figure 3B); this regression was
stronger in HCV-negative patients (n ⫽ 180; R ⫽ 0.53)
compared with HCV-positive patients (n ⫽ 30; R ⫽ 0.15,
data not shown).
Effect of Deferasirox on Liver Iron Burden
and Serum Ferritin Levels
For all patients in this subanalysis (n ⫽ 219), mean
LIC decreased from 15.7 ⫾ 9.9 mg Fe/g dw at baseline to

Figure 3. (A) Mean absolute change in ALT by EOS and (B) scatter plot of absolute changes from treatment initiation for LIC and ALT. Only patients
with LIC and ALT at both baseline and EOS are included. ⴱP is significant at EOS compared with baseline. †P is nonsignificant at EOS compared with
baseline.
October 2011
Figure 4. (A) Mean absolute change in liver iron concentration ⫾ standard deviation (SD) from baseline to EOS; (B) median absolute change in serum
ferritin levels ⫾ 25th and 75th percentiles from baseline to EOS; mean absolute change in (C) TIS ⫾ SD; and (D) liver iron ratio ⫾ SD from baseline to
EOS. ⴱP is significant at EOS compared with baseline. †P is nonsignificant at EOS compared with baseline.
10.1 ⫾ 8.2 mg Fe/g dw at EOS (mean absolute change of
⫺5.5 ⫾ 10.6 mg Fe/g dw; P ⬍ .001; mean relative change
of ⫺15.1% ⫾ 93.4%; P ⫽ .020). For patients with LIC
assessments at both baseline and EOS (n ⫽ 210), overall
LIC response rate was 63.8% (134 of 210). Mean reduc-
tions of ⫺10.7 ⫾ 8.2 mg Fe/g dw (P ⬍ .001) (mean relative
change of ⫺57.1% ⫾ 24.1%; P ⬍ .001) from baseline of
18.3 mg Fe/g dw were observed in Group A and mean
increases of 3.9 ⫾ 7.4 mg Fe/g dw (P ⬍ .001) (mean
relative change of 59.0% ⫾ 120.6%; P ⬍ .001) from base-
line of 10.7 mg Fe/g dw were observed in Group B (Figure
4A). Mean LIC decreased regardless of HCV status (HCV-
positive: mean absolute change of ⫺4.9 ⫾ 10.0; P ⫽ .012;
mean relative change of 1.1% ⫾ 112.0%; P ⫽ .958; HCV-
negative: mean absolute change of ⫺5.6 ⫾ 10.7; P ⬍ .001;
mean relative change of ⫺17.8% ⫾ 90.0%; P ⫽ .009; Figure
4A). In all patients, median serum ferritin levels decreased
from 2069 ng/mL at baseline to 1111 ng/mL at EOS
(median absolute change of ⫺669 ng/mL; P ⬍ .001).
Median absolute changes of ⫺1149 ng/mL (P ⬍ .001) and
⫹71.8 ng/mL (P ⫽ .64) were observed in Group A and B,
respectively (Figure 4B). Serum ferritin levels decreased in
the study in both HCV-positive and HCV-negative pa-
tients (median absolute changes of ⫺439 ng/mL; P ⫽ .043
and ⫺695 ng/mL; P ⬍ .001, respectively; Figure 4A). In all
patients, mean TIS decreased from 25.1 ⫾ 11.0 at baseline
IMPROVEMENT IN LIVER PATHOLOGY WITH DEFERASIROX 1207
CLINICAL LIVER
to 17.0 ⫾ 10.8 at EOS (mean absolute change of ⫺8.2 ⫾
13.3; P ⬍ .001). Mean decreases in TIS were only observed
in Group A (mean absolute change of ⫺13.7 ⫾ 10.9; P ⬍
.001; Group B: ⫹1.5 ⫾ 11.2; P ⫽ .269; Figure 4C). Signif-
icant reductions in TIS were observed regardless of HCV
status (HCV-positive patients: ⫺6.1 ⫾ 13.4; P ⫽ .027;
HCV-negative patients: ⫺8.5 ⫾ 13.2; P ⬍ .001; Figure 4C).
In all patients, mean liver iron ratio remained stable, from
66.5% ⫾ 14.8% at baseline to 64.2% ⫾ 26.6% at EOS (mean
absolute change of ⫺2.1% ⫾ 27.3%; P ⫽ .276). Nonsignif-
icant reductions were observed in HCV-positive (mean
absolute change of ⫺5.7% ⫾ 28.0%; P ⫽ .318) and HCV-
negative (mean absolute change of ⫺1.7% ⫾ 27.3%; P ⫽
.435) patients (Figure 4D). A summary table describing
absolute change from baseline until EOS for all efficacy
outputs is shown in Table 3.
Discussion
The current subanalysis demonstrates that treat-
ment with deferasirox for at least 3 years was associated
with stability or improvement in liver fibrosis and histo-
logical necroinflammation, as well as reduced ALT levels
in iron-overloaded patients with ␤-thalassemia. This im-
provement was observed in patients who met the LIC
response criteria and by those who did not, and occurred
 1208 DEUGNIER ET AL
Table 3. Summary of Changes in Liver and Iron Parameters From Baseline Until EOS
Group A
Summary of results (n ⫽ 134)
Absolute change from baseline
Mean necroinflammatory score (⫾SD) ⫺1.5 ⫾ 1.8
Mean serum ALT, IU/mL (⫾SD) ⫺21.5 ⫾ 35.0
Mean LIC, mg Fe/g dw (⫾SD) ⫺10.7 ⫾ 8.2
Mean total iron score (⫾SD) ⫺13.7 ⫾ 10.9
Mean liver iron ratio, % (⫾SD) ⫺2.4 ⫾ 26.1
Median serum ferritin, ng/mL (range) ⫺1149 (⫺10,164 to 1681)
SD, standard deviation.
independently of evidence of previous HCV exposure (as
assessed by the presence of anti-HCV antibodies). To our
CLINICAL LIVER
knowledge, this is the first analysis that demonstrates
regression of fibrosis in patients with ␤-thalassemia dur-
ing iron chelation therapy.
Overall improvement in liver fibrosis and necroin-
flammation with deferasirox treatment seen in this
study is striking compared with other reports in the
literature. Indeed, there are limited data supporting the
ability of any therapeutic agent to actively reverse liver
fibrosis.8 –12 In a study of 36 patients with genetic
hemochromatosis, regression of fibrosis (defined using
the METAVIR grading system25) was observed in almost
half of patients following venesection therapy (phlebot-
omy) after a mean of 9.7 ⫾ 5.2 years.26 Regression of
cirrhosis has also been observed in patients with
␤-thalassemia after bone marrow transplantation.27 In
this study, the higher proportion of worsening fibrosis
seen in the deferasirox cohort vs that in the crossover
cohort may be influenced by the relative underdosing of
deferasirox during the first year in patients who re-
ceived the drug from the outset. However, the percent-
age of patients who improved or stabilized remained
relatively similar between the crossover and deferasirox
cohorts. Of the 22 patients who demonstrated worsen-
ing fibrosis, only 12 had a baseline LIC of ⱖ15 mg Fe/g
dw; while 7 patients had a baseline LIC of ⱕ10 mg Fe/g
dw. When comparing directly to other iron chelators,
the results presented here demonstrate more favorable
results. Small studies using deferiprone (n ⫽ 17⫺56)
have demonstrated no significant change in fibrosis
score in patients with ␤-thalassemia treated for approx-
imately 3 years.6,7 Other studies have demonstrated
that deferiprone treatment is associated with the pro-
gression of fibrosis in patients with ␤-thalassemia.28 An
early study with DFO in 9 patients with ␤-thalassemia
showed stabilization of hepatic fibrosis during treat-
ment over a 7-year period.13 A more recent study in a
larger cohort of patients (n ⫽ 32, mostly thalassemia
major) treated with DFO for at least 2 years demon-
strated stabilization of fibrosis, with an annual rate of
progression of 0.13 ⫾ 0.50.5 The concept of the free
radical scavenging ability of DFO has been demon-
strated,29 and the potential contribution of such an
ability to influence fibrosis progression may be of in-
GASTROENTEROLOGY Vol. 141, No. 4
Group B Total
(n ⫽ 76) (n ⫽ 210)
–0.8 ⫾ 1.6 ⫺1.2 ⫾ 1.8
6.8 ⫾ 25.2 ⫺11.3 ⫾ 34.6
3.9 ⫾ 7.4 ⫺5.5 ⫾ 10.6
1.5 ⫾ 11.2 ⫺8.2 ⫾ 13.2
⫺1.9 ⫾ 29.6 ⫺2.2 ⫾ 27.4
72 (⫺2735 to 4929) ⫺675 (⫺10,164 to 4929)
terest in future studies. Although previous studies sup-
port the concept that other iron chelators might lead to
stabilization of hepatic fibrosis with long-term treat-
ment, the results presented here show that deferasirox
can lead to regression of fibrosis, underscoring the
importance of these results. The stability or improve-
ment in liver fibrosis and necroinflammation was ob-
served in patients who met the LIC response criteria as
specified in this study and those who did not, suggest-
ing the observed effects may be independent of the
drug’s iron removal effects. Although deferasirox is
likely to bind available circulating iron such as labile
plasma iron (LPI), in many patients a significant im-
provement is seen in fibrosis with no change, or even
with increasing amounts of iron in the liver. This result
is corroborated by the lack of a correlation between
reductions in LIC and reductions in Ishak staging/
grading in this study. Although reductions in non-
transferrin-bound iron such as LPI (a redox-active sub-
fraction of plasma non-transferrin-bound iron) do not
always correlate with reduced LIC levels,30 iron-induced
oxidative damage can lead to fibrosis.31 Furthermore,
deferasirox treatment has been shown to provide sus-
tained suppression of LPI.32
Although not analyzed in this study, an assessment of
LPI levels among patients with regression of fibrosis may
be of interest for future studies. It is interesting that
Angelucci and colleagues noted that the risk of fibrosis
progression correlated with hepatic iron content in pa-
tients cured of ␤-thalassemia with bone marrow trans-
plantation.4 The data we present here, although indicat-
ing a lack of correlation between change in LIC with the
change in Ishak staging, do not assess the effect of abso-
lute values of LIC on fibrosis changes. In their study,
Angelucci and colleagues, when assessing the effect of LIC
on progression of fibrosis in successfully transplanted
thalassemia patients, found a threshold effect at an LIC of
approximately 16 mg Fe/g dw, below which patients did
not experience progression of fibrosis.4 Furthermore, ele-
vated LIC ⬎300 ␮M/g (approximately 16.75 mg Fe/g dw)
was also associated with an increased risk of abnormal
ALT values in iron-overloaded patients with myelodys-
plastic syndromes.33
These observations raise new questions regarding the
mechanism by which deferasirox leads to improvement
October 2011
in liver fibrosis. Our results suggest this mechanism
may not be a secondary effect of simply reducing he-
patic iron burden. This effect might be due to the fact
that deferasirox is concentrated and excreted via the
biliary system,34 with the potential of intrahepatocytic
trafficking, perhaps facilitating a specific pharmacolog-
ical effect of deferasirox or its metabolites on the
pathophysiological mechanisms that modulate fibrosis.
Recent studies have shown an inhibitory effect of de-
ferasirox on the transcriptional nuclear factor ␬B,35 a
protein that has also been shown to be implicated in
the development of fibrosis in the lung.36 Further mo-
lecular biology studies are required to explore these
hypotheses. We cannot formally exclude, however, the
possibility that decreases in LIC did not accurately
reflect total body iron clearance or the possibility that
patients in Group B experienced more pronounced iron
reduction in extrahepatic sites. Nonetheless, these re-
sults are encouraging and warrant further studies to
investigate the potential effects of deferasirox in pre-
venting iron-induced tissue fibrosis in organs other
than the liver, such as the endocrine organs or the
heart. In support of this concept, recent preclinical data
in which deferasirox treatment was administered to an
iron-overloaded gerbil model were associated with at-
tenuated cardiac fibrosis.37
Consistent with multiple previous studies38 – 40 that
have demonstrated the effects of deferasirox in reduc-
ing iron burden, in this subanalysis deferasirox treat-
ment for at least 3 years led to improvement in markers
of iron loading, including LIC, serum ferritin levels,
and total iron score. LIC values ⬎15 mg Fe/g dw are
associated with a greatly increased risk of iron-induced
cardiac disease and early death22; in this analysis, LIC
decreased from 15.7 ⫾ 9.9 mg Fe/g dw at baseline to
10.1 ⫾ 8.2 mg Fe g/dw at EOS. Response criteria in this
study were based on LIC measurements, one of the
most robust markers available to assess iron load. Se-
rum ferritin assessment, although cheap and conve-
nient for serial measurements, can be affected by in-
flammation, infection, and the presence of liver disease.
Response to iron chelation was independent of evidence
of HCV exposure, suggesting that previous exposure to
the virus does not lead to diminished response to de-
ferasirox therapy, a result consistent with a subanalysis
from the large Evaluation of Patients’ Iron Chelation
(EPIC) study.41 It should be noted that a previous
analysis of data from studies 107 and 108 demon-
strated improvements in hepatocellular inflammation
and liver function after treatment with deferasirox/
DFO for 1 year.18 Treatment during this time frame,
however, did not lead to improvements in liver fibrosis,
suggesting this effect requires longer-term drug expo-
sure, a result that is confirmed by the current findings.
Reductions in both LIC and serum ferritin levels mir-
rored reductions in ALT, an important indicator of hep-
atocellular damage. Regression analysis revealed a corre-
lation between reduction in ALT and reductions in LIC,
IMPROVEMENT IN LIVER PATHOLOGY WITH DEFERASIROX 1209
suggesting a positive impact of deferasirox on liver func-
tion. This is consistent with previous studies that have
demonstrated an association between elevated ALT and
LIC in patients with myelodysplastic syndromes33 and an
association between decreasing serum ferritin and ALT in
myelodysplastic syndromes42,43 and in a variety of trans-
fusion-dependent anemias.44
Because of the requirement for paired liver biopsies
and LIC assessments after at least 3 years of deferasirox
treatment, only 210 patients with ␤-thalassemia from
671 enrolled in the core studies were included in this
subanalysis. The study also has a number of other
limitations, including the small number of patients
with evidence of HCV exposure, the lack of availability
of HCV RNA data or of HCV status assessment during
CLINICAL LIVER
the treatment period, as well as the variable duration of
deferasirox exposure. A small number of patients
(⬍10%) in the 107E and 108E studies also received
antiviral medication (⬎90% of these patients received
nucleosides and nucleotides excluding reverse tran-
scriptase inhibitors). Although liver biopsy samples
from both studies were analyzed in a single central
laboratory, methodological limitations of the subanaly-
sis include the variable size and quality of biopsy sam-
ples obtained.
Conclusions
Data presented here suggest that long-term defera-
sirox treatment is associated with stability or improve-
ment of liver fibrosis and necroinflammation in heavily
iron-overloaded patients with ␤-thalassemia, an effect
that was independent of reductions in LIC as defined by
the response criteria or evidence of previous HCV expo-
sure.
Supplementary Material
Note: To access the supplementary material
accompanying this article, visit the online version of
Gastroenterology at www.gastrojournal.org, and at doi:
10.1053/j.gastro.2011.06.065.
References
1. Andrews NC. Disorders of iron metabolism. N Engl J Med 1999;
341:1986 –1995.
2. Kushner JP, Porter JP, Olivieri NF. Secondary iron overload. Hema-
tology Am Soc Hematol Educ Program 2001;47– 61.
3. Ardalan FA, Osquei MR, Toosi MN, et al. Synergic effect of chronic
hepatitis C infection and beta thalassemia major with marked
hepatic iron overload on liver fibrosis: a retrospective cross-sec-
tional study. BMC Gastroenterol 2004;4:17.
4. Angelucci E, Muretto P, Nicolucci A, et al. Effects of iron overload
and hepatitis C virus positivity in determining progression of liver
fibrosis in thalassemia following bone marrow transplantation.
Blood 2002;100:17–21.
5. Berdoukas V, Bohane T, Tobias V, et al. Liver iron concentration
and fibrosis in a cohort of transfusion-dependent patients on
long-term desferrioxamine therapy. Hematol J 2005;5:572–578.
6. Wanless IR, Sweeney G, Dhillon AP, et al. Lack of progressive
hepatic fibrosis during long-term therapy with deferiprone in sub-
 1210 DEUGNIER ET AL
jects with transfusion-dependent beta-thalassemia. Blood
2002;100:1566 –1569.
7. Wu SF, Peng CT, Wu KH, et al. Liver fibrosis and iron levels during
long-term deferiprone treatment of thalassemia major patients.
Hemoglobin 2006;30:215–218.
8. Farci P, Roskams T, Chessa L, et al. Long-term benefit of inter-
feron alpha therapy of chronic hepatitis D: regression of advanced
hepatic fibrosis. Gastroenterology 2004;126:1740 –1749.
9. Poynard T, Bedossa P, Opolon P. Natural history of liver fibrosis
progression in patients with chronic hepatitis C. The OBSVIRC,
METAVIR, CLINIVIR, and DOSVIRC groups. Lancet 1997;349:825–
832.
10. Shiratori Y, Imazeki F, Moriyama M, et al. Histologic improvement
of fibrosis in patients with hepatitis C who have sustained re-
sponse to interferon therapy. Ann Intern Med 2000;132:517–
524.
11. Das K, Chowdhury A. Hepatic fibrosis: can we treat it clinically?
CLINICAL LIVER
Trop Gastroenterol 2008;29:76 – 83.
12. Povero D, Busletta C, Novo E, et al. Liver fibrosis: a dynamic and
potentially reversible process. Histol Histopathol 2010;25:1075–
1091.
13. Barry M, Flynn DM, Letsky EA, et al. Long-term chelation therapy in
thalassaemia major: effect on liver iron concentration, liver histol-
ogy, and clinical progress. Br Med J 1974;2:16 –20.
14. Cappellini MD, Cohen A, Piga A, et al. A phase 3 study of defera-
sirox (ICL670), a once-daily oral iron chelator, in patients with
␤-thalassemia. Blood 2006;107:3455–3462.
15. Porter J, Galanello R, Saglio G, et al. Relative response of patients
with myelodysplastic syndromes and other transfusion-dependent
anaemias to deferasirox (ICL670): a 1-yr prospective study. Eur J
Haematol 2008;80:168 –176.
16. Cappellini MD, Perrotta S, Agaoglu L, et al. Efficacy and safety of
deferasirox (Exjade®) in patients with ␤-thalassemia major treated
for up to 5 years. Blood 2009;114(22):abst 4063.
17. Aydinok Y, Agaoglu L, Bejaoui M, et al. Growth and develop-
ment of pediatric patients with beta-thalassaemia treated with
deferasirox for up to 5 years. Haematologica 2010;95(Suppl
2):abst 1040.
18. Brissot P, Turlin B, Forni G-L, et al. Iron chelation therapy with
deferasirox (Exjade®, ICL670) or deferoxamine results in reduced
hepatocellular inflammation and improved liver function in pa-
tients with transfusion-dependent anemia. Blood 2005;106(11):
abst 823.
19. Barry M, Sherlock S. Measurement of liver-iron concentration in
needle-biopsy specimens. Lancet 1971;1:100 –103.
20. Angelucci E, Brittenham GM, McLaren CE, et al. Hepatic iron
concentration and total body iron stores in thalassemia major.
N Engl J Med 2000;343:327–331.
21. Soriano-Cubells MJ, Aparisi-Querada L. Rapid determination of
copper, iron and zinc in liver biopsies. Atomic Spectrosc 1984;5:
217–222.
22. Olivieri NF, Brittenham GM. Iron-chelating therapy and the treat-
ment of thalassemia. Blood 1997;89:739 –761.
23. Deugnier YM, Loreal O, Turlin B, et al. Liver pathology in genetic
hemochromatosis: a review of 135 homozygous cases and their
bioclinical correlations. Gastroenterology 1992;102:2050 –2059.
24. Ishak K, Baptista A, Bianchi L, et al. Histological grading and
staging of chronic hepatitis. J Hepatol 1995;22:696 – 699.
25. Bedossa P, Poynard T. An algorithm for the grading of activity in
chronic hepatitis C. The METAVIR Cooperative Study Group. Hepa-
tology 1996;24:289 –293.
26. Falize L, Guillygomarc’h A, Perrin M, et al. Reversibility of hepatic
fibrosis in treated genetic hemochromatosis: a study of 36 cases.
Hepatology 2006;44:472– 477.
27. Muretto P, Angelucci E, Lucarelli G. Reversibility of cirrhosis in
patients cured of thalassemia by bone marrow transplantation.
Ann Intern Med 2002;136:667– 672.
GASTROENTEROLOGY Vol. 141, No. 4
28. Olivieri NF, Brittenham GM, McLaren CE, et al. Long-term safety
and effectiveness of iron-chelation therapy with deferiprone for
thalassemia major. N Engl J Med 1998;339:417– 423.
29. Zhu BZ, Har-El R, Kitrossky N, et al. New modes of action of
desferrioxamine: scavenging of semiquinone radical and stimula-
tion of hydrolysis of tetrachlorohydroquinone. Free Radic Biol Med
1998;24:360 –369.
30. Greenberg PL, Koller CA, Cabantchik ZI, et al. Prospective assess-
ment of effects on iron-overload parameters of deferasirox therapy
in patients with myelodysplastic syndromes. Leuk Res 2010;34:
1560 –1565.
31. Porter J. Pathophysiology of iron overload. Hematol Oncol Clin
North Am 2005;19(Suppl 1):7–12.
32. List AF, Baer MR, Steensma D, et al. Deferasirox (ICL670; Ex-
jade®) reduces serum ferritin (SF) and labile plasma iron (LPI) in
patients with myelodysplastic syndromes (MDS). Blood 2007;
110(11):abst 1470.
33. Jensen PD, Jensen FT, Christensen T, et al. Relationship between
hepatocellular injury and transfusional iron overload prior to and
during iron chelation with desferrioxamine: a study in adult pa-
tients with acquired anemias. Blood 2003;101:91–96.
34. Bruin GJ, Faller T, Wiegand H, et al. Pharmacokinetics, distribu-
tion, metabolism and excretion of deferasirox and its iron complex
in rats. Drug Metab Dispos 2008;36:2523–2538.
35. Messa E, Carturan S, Maffé C, et al. Deferasirox is a powerful NF-kB
inhibitor in myelodysplastic cells and in leukemia cell lines acting
independently from intracellular iron chelation and reactive oxygen
species scavenging. Haematologica 2010;95:1308 –1316.
36. Fujimoto H, D’Alessandro-Gabazza CN, Palanki MS, et al. Inhibi-
tion of nuclear factor-kappaB in T cells suppresses lung fibrosis.
Am J Respir Crit Care Med 2007;176:1251–1260.
37. Wang Y, Wu M, AL-Rousan R, et al. Iron-induced cardiac damage:
role of apoptosis and deferasirox intervention. J Pharmacol Exp
Ther 2011;336:56 – 63.
38. Cappellini MD, Porter JB, El-Beshlawy A, et al. Tailoring iron che-
lation by iron intake and serum ferritin trends: the prospective
multicenter EPIC study of deferasirox in 1744 patients with vari-
ous transfusion-dependent anemias. Haematologica 2010;95:
557–566.
39. Taher A, El-Beshlawy A, Elalfy MS, et al. Efficacy and safety of
deferasirox, an oral iron chelator, in heavily iron-overloaded pa-
tients with ␤-thalassaemia: the ESCALATOR study. Eur J Haematol
2009;82:458 – 465.
40. Vichinsky E, Coates T, Thompson AA, et al. Deferasirox (Exjade®),
the once-daily oral iron chelator, demonstrates safety and efficacy
in patients with sickle cell disease (SCD): 3.5-year follow-up.
Blood 2008;112(11):abst 1420.
41. Cappellini MD, Elalfy M, El-Beshlawy A, et al. Efficacy of deferasirox in
patients with thalassaemia major is not affected by a history of
hepatitis B or C. Haematogica 2010;95(Suppl 2):abst 1860.
42. Gattermann N, Finelli C, Della Porta M, et al. Deferasirox in
iron-overloaded patients with transfusion-dependent myelodys-
plastic syndromes: results from the large 1-year EPIC study. Leuk
Res 2010;34:1143–1150.
43. Wimazal F, Nösslinger T, Baumgartner C, et al. Deferasirox in-
duces regression of iron overload in patients with myelodysplastic
syndromes. Eur J Clin Invest 2009;39:406 – 411.
44. Brissot P, Deugnier Y, Cianciulli P, et al. Dose-dependent reduc-
tions in serum transaminases and ferritin levels correlate during
long-term therapy with deferasirox (Exjade, ICL670). Book of Ab-
stracts International BioIron Society 2007;abst 105.
Received March 18, 2011. Accepted June 9, 2011.
Reprint requests
Address requests for reprints to: Yves Deugnier, MD, Liver Disease
Unit and Inserm U-991, University Hospital Pontchaillou, Rennes,
France. e-mail: yves.deugnier@univ-rennes1.fr; fax: ⴙ33 2 9928
4112.
October 2011
Acknowledgments
We thank John Mc Intyre for medical editorial assistance with this
manuscript.
Conflicts of interest
These authors disclose the following: Dr Deugnier received
honoraria from Novartis. Dr Cappellini is a member of the
Speakers’ Bureau at Novartis and received lecture fees from
Novartis. Dr Porter received consulting fees, research grant
IMPROVEMENT IN LIVER PATHOLOGY WITH DEFERASIROX 1211
funding, and lecture fees from Novartis. Dr Brissot received
honoraria and research funding from Novartis. Drs Giannone,
Zhang, and Griffel are employees of Novartis. The remaining
authors disclose no conflicts.
Funding
Financial support for medical editorial assistance was provided by
Novartis Pharmaceuticals. This study was sponsored by Novartis
Pharma AG.
CLINICAL LIVER
1211.e1 DEUGNIER ET AL
Appendix
The following centers and investigators partici-
pated in the 107 and 108 studies:
Study 107
Argentina: G. Drelichman, Hospital de Ninos Ri-
cardo Gutierrez, Buenos Aires; N. Watman, Hospital
Ramos Mejía, Buenos Aires; A. Berreta, Hospital Privado
de Cordoba, Cordoba.
Belgium: C. Vermylen, Cliniques Universtaires St Luc,
Bruxelles; D. Boulet, CHR St Joseph, Mons; A. Ferster,
H.U.D.E. Reine Fabiola, Laeken; M.-F. Dresse, CHR Cita-
delle, Liege.
Brazil: M. Verissimo, V. Pereira, Centro Infantil de
Investigacoes Hematologicas Dr Domingos A. Boldrini,
Campinas; S. Loggetto, M. L. Silva, Centro de Hemato-
logia Sao Paulo, Sao Paulo.
Canada: N Olivieri, Toronto General Hospital, To-
ronto, Ontario; S Abish, Montreal Children’s Hospital,
Montreal, Quebec.
France: I. Thuret, Hopital de la Timone Enfants, Mar-
seille; M. De Montalembert, Hopital Necker, Paris; D.
Bachir, CHU Henri Mondor, Creteil; R. Girot, Hôpital
Tenon, Paris; and G. Salles, Centre Hospitalier Lyon-Sud,
Pierre-Benite.
Germany: G. Janka-Schaub, Universitaetskrankenhaus
Eppendorf, Hamburg;
E. Kohne, Universitaetsklinikum Ulm, Ulm; G. Janssen,
Universitaetsklinikum Düsseldorf, Düsseldorf; T. Kling-
biel, Universitaetsklinikum Frankfurt, Frankfurt; and G.
Strauss, Universitaetskinderklinik, Berlin.
Greece: C. Kattamis, V. Ladis, and H. Berdoussi, “Agia
Sofia” Children’s Hospital, First Department of Pediat-
rics, Athens University School of Medicine, Athens; A.
Koussi and I. Tsatra, “Hippokration” Hospital of Thes-
saloniki, First Department of Pediatrics, Hematology Di-
vision, “Aristotle” University of Thessaloniki, Thessalo-
niki; N. Zoumbos and I. Constandinidou, University
Hospital of Patras, Department of Internal Medicine,
Haematology Division, Patras; K. Bourantas, University
Hospital of Ioannina, Haematology Department, Ioan-
nina.
Italy: S. Perrotta, D. Di Pinto, B. Nobili, Policlinico, II
Università di Napoli, Naples; M.D. Cappellini and L.
Zanaboni, Università di Milano, Ca Granda Foundation
IRCCS, Milan; M. Capra, A. Montesanto, Ospedale Civico
e Benefratelli G. di Cristina M. Ascoli, Palermo; R. Origa,
A. Zappu, Ospedale Regionale Microcitemie—A.S.L.n. 8,
Cagliari; C. Magnano, S. Anastasi, Az. Osp. di Rilievo
Nazionale e di Alta Spec. Garibaldi, Catania; A. Mangia-
gli, S. Campise, Ospedale Umberto I, Siracusa; G. Masera,
N Masera, Ospedale Nuovo San Gerardo Università degli
Studi di Milano, Monza; V. De Sanctis, MR Gamberini,
Az. Osp. Universitaria Sant’Anna, Ferrara; P. Cianciulli, F.
Sorrentino, Ospedale S. Eugenio, Rome; D. Gallisai, Isti-
tuto di Clinica Pediatrica Amerigo Filia, Sassari; G.
GASTROENTEROLOGY Vol. 141, No. 4
Quarta, Ospedale A. Perrino—Az. USL BR1, Brindisi; A.
Saviano, Azienda Ospedaliera di Rilievo Nazinale A
Cardarelli, Naples; S. Cossu, Ospedale Civile SS. Annun-
ziata—A.U.S.L. N. 1 Sassari, Sassari; ME Lai, Ospedale
Regionale Microcitemie—A.S.L.n. 8, Cagliari; A Maggio,
Azienda Ospedaliera V. Cervello, Palermo; R. Malizia,
Azienda Ospedaliera Villa Sofia—CTO, Palermo; A. Piga,
Universitá di Torino, Turin; G.L. Forni, Ospedale Galli-
era, Genoa; F. Locatelli, Policlinico San Matteo—IRCCS,
Pavia; C. Borgna-Pignatti, Az. Osp. Universitaria
Sant’Anna—Università degli Studi, Ferrara.
Tunisia: M. Bejaoui, Centre National des Greffes de la
Moelle Osseuse, Tunis; S. Fattoum, Hopital d’enfants,
Tunis; B. Meddeb, Hôpital Aziza Othmana, Tunis.
Turkey: L. Agaoglu, Z Karakas, Istanbul University
Istanbul Medical Faculty, Istanbul; Y. Aydinok, Ege Uni-
versity Medical Faculty, Izmir; Y. Kilinc, I. Sasmaz, Cu-
kurova University Medical Faculty, Adana; D. Canatan,
Suleyman Demirel University Medical Faculty, Isparta; Z.
Uysal, M. Ertem, Ankara University Medical Faculty, An-
kara.
United Kingdom: J. Porter, P. Eleftheriou-Kokkinos,
University College Hospital, London.
United States: P. Giardina, New York Presbyterian Hos-
pital, New York, NY; A. Cohen, Children’s Hospital of
Philadelphia, Philadelphia, PA; T. Coates, Children’s Hos-
pital Los Angeles, Los Angeles, CA; A. Thompson, Chil-
dren’s Memorial Hospital, Chicago, IL; E. Vichinsky,
Children’s Hospital and Research Center at Oakland,
Oakland, CA; E. Neufeld, Boston Children’s Hospital,
Boston, MA; M. Jeng, Stanford Hospital, Stanford, CA.
Study 108
Belgium: J. Maertens, M. Delforge, U.Z. Gasthuis-
berg, Leuven; L. De Busscher, CHU Tivoli, La Louvière;
Selleslag D, A.Z. Sint-Jan, Brugge; C. Vermylen, Cliniques
Universtaires St. Luc, Bruxelles; Noens L, Universitair
Ziekenhuis Gent, Gent; P. Zachee, Algemeen Centrum
Ziekenhuis Antwerpen A.Z. Stuivenberg, Antwerpen.
Canada: N. Olivieri, Toronto General Hospital, To-
ronto; D. Soulieres, CHUM—Pavillion Notre Dame, Mon-
treal; S Abish, Montreal Children’s Hospital, Montreal.
France: C. Rose, N. Cambier, Centre Hospitalier Saint-
Vincent, Lille Cédex; G. Tchernia, Hopital Kremlin Bice-
tre, Le Kremlin Bicetre; G. Dine, Hopital Les Hauts Clos,
Troyes; D. Bachir, Hopital Henri Mondor, Creteil.
Germany: N. Gattermann, F. Fox, Universitaetsklini-
kum Düsseldorf, Düsseldorf; A. Ganser, M. Stadler, Me-
dizinische Hochschule, Hannover, Hannover; H. Cario,
Kohne E, Universitaetsklinikum Ulm, Ulm.
Italy: M.D. Cappellini, L. Zanaboni, Universitá di Mi-
lano, Ca Granda Foundation IRCCS, Milano; A. Piga, G.
Militerno, G. Donato, Ospedale Regina Margherita,
Torino; G. Quarta, A. Melpignano, Ospedale A. Perrino—
Az. USL BR1, Brindisi; G. Forni, A. Lavagetto, Ospedale
Galliera, Genova; R. Galanello, GB Leoni, Ospedale Re-
October 2011
gionale Microcitemie—A.S.L.n. 8, Cagliari; G. Saglio, E.
Messa, Azienda Sanitaria Ospedaliera S. Luigi Gonzaga,
Orbassano; M. Cazzola, Policlinico S. Matteo—IRCCS -
Università degli Studi, Pavia; G. Alimena, I. Carmosino,
Azienda Policlinico Umberto I—Università La Sapienza,
Roma; M. Baccarani, Policlinico S. Orsola—Malpighi—Az.
Osp. di Bologna, Bologna; E. Angelucci, Ospedale Onco-
logico A. Businco—A.S.L. n. 8, Cagliari; P. Cianciulli,
Ospedale S. Eugenio, Roma; Morra E, Az. Osp. Niguarda
Ca’ Granda, Milano.
IMPROVEMENT IN LIVER PATHOLOGY WITH DEFERASIROX 1211.e2
UK: J. Porter, P. Eleftheriou, M. Cummins, University
College Hospital, London; G. Mufti, King’s College Hos-
pital, London.
USA: E. Vichinsky, P. Harmatz, S. Singer, Children’s
Hospital & Research Center at Oakland, Oakland, CA; M.
Cunningham, E. Neufeld, Children’s Hospital Boston,
Boston, MA; P. Giardina, New York Presbyterian Hospi-
tal, New York, NY; M. Jeng, P. Greenberg, Stanford Hos-
pital, Stanford, CA; A. Cohen, J. Kwiatkowski, C. Manno,
Children’s Hospital of Philadelphia, Philadelphia, PA.
1211.e3 DEUGNIER ET AL GASTROENTEROLOGY Vol. 141, No. 4

Supplementary Table 1. Changes in Ishak Staging From Start of Deferasirox Treatment to EOS in the Deferasirox and
Crossover Cohorts
Change in Ishak stage from start of treatment to EOS, n (%)

Group LIC response status ⱕ2 ⫺1, 0, ⫹1 ⱕ2 Missing Total

Deferasirox (n ⫽ 125) Group A 19 (23.2) 53 (64.6) 6 (7.3) 4 (4.9) 82 (100)
Group B 10 (26.3) 17 (44.7) 9 (23.7) 2 (5.3) 38 (100)
Missing 0 (0.0) 4 (80.0) 0 (0.0) 1 (20.0) 5 (100)
Total 29 (23.2) 74 (59.2) 15 (12.0) 7 (5.6) 125 (100)
Crossover (n ⫽ 94) Group A 16 (30.8) 27 (51.9) 2 (3.8) 7 (13.5) 52 (100)
Group B 13 (34.2) 20 (52.6) 3 (7.9) 2 (5.3) 38 (100)
Missing 1 (25.0) 1 (25.0) 2 (50.0) 0 (0.0) 4 (100)
Total 30 (31.9) 48 (51.1) 7 (7.4) 9 (9.6) 94 (100)
Total (n ⫽ 219) Group A 35 (26.1) 80 (59.7) 8 (6.0) 11 (8.2) 134 (100)
Group B 23 (30.3) 37 (48.7) 12 (15.8) 4 (5.3) 76 (100)
Missing 1 (11.1) 5 (55.6) 2 (22.2) 1 (11.1) 9 (100)
Total 59 (26.9) 122 (55.7) 22 (10.0) 16 (7.3) 219 (100)