determination of deferasirox plasma concentrations: do gender, physical and genetic differences...
TRANSCRIPT
Acc
epte
d A
rtic
le
This article has been accepted for publication and undergone full peer review but has not been through the copyediting, typesetting, pagination and proofreading process, which may lead to differences between this version and the Version of Record. Please cite this article as doi: 10.1111/ejh.12419
This article is protected by copyright. All rights reserved.
Article Type: Original Article
Accepted date: 24-Jul-2014
Title: Determination of deferasirox plasma concentrations: do gender, physical and genetic
differences affect chelation efficacy?
Authors: F Mattiolia, M Puntonib, V Marinia, C Fucilea, G Milanoa, L Robbianoa, S Perrottac, V. Pintod, A
Martellia, G L Fornid
Affiliation:
aDepartment of Internal Medicine, Clinical Pharmacology and Toxicology Unit, University of Genoa,
Genoa, Italy
bClinical Trial Unit, Scientific directorate, E.O. Galliera, Genoa, Italy
cDepartment of Pediatrics, Second University of Naples, Naples, Italy
dSSD Ematologia – Centro della Microcitemia E.O. Galliera, Genoa, Italy
Running Title: Gender and physical differences affect deferasirox
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
Corresponding author: Prof. Francesca Mattioli
aDepartment of Internal Medicine, Clinical Pharmacology and Toxicology Unit, University of Genoa,
Genoa, Italy
Address: Viale Benedetto XV, n. 2. I-16132 Genoa, Italy
Phone: +390103538850; Fax: +390103538232
Email: [email protected]
Abstract
Objectives. Bioavailability of deferasirox (DFX) is significantly affected by timing of administration
relative to times and to composition of meals. Its elimination half-life is also highly variable – in some
patients as a result of gene polymorphisms. Understanding if deferasirox plasma levels are related to
specific characteristics of patients could help physicians to devise a drug regimen tailored the
individual patient.
Methods. We analysed deferasirox plasma concentrations (CDFX) in 80 patients with transfusion-
dependent anemias, such as thalassemia, by a HPLC assay. We used a multivariate linear regression
model to find significant associations between CDFX and main patients clinical/demographical
characteristics of patients. All patients were genotyped for UGT1A1.
Results. Fifty-six patients were female and 24 were male, the great majority (88%) affected by β-
thalassemia, 15 were children and adolescents. No statistical correlation was detectable between
CDFX and DFX dose (p=0.6). Age, time from last drug intake to blood sampling and ferritin levels in the
6 months before study initiation were significantly and inversely associated with CDFX in univariate
analysis. In the multivariate analysis the only two factors independently and inversely associated with
CDFX levels were time from last drug intake to blood sampling and ferritin levels (p=0.006). A
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
significant inverse correlation (p=0.03) was observed between CDFX and UGT1A1*28 gene
polymorphism, but only in patients with levels of lean body mass (LBM) below the median (p for
interaction=0.05).
Conclusions. The results could indicate that a higher plasma DFX concentration could be
associated with greater chelation efficacy. Since a correlation between dose and CDFX was not
demonstrated, it seems useful to monitor the concentrations to optimize and determine the
most appropriate dose for each patient. Interesting results emerged from the analysis of
genetic and physicals characteristics of patients: LBM was a borderline significant effect
modifier of the relationship between UGT1A1 polymorphisms and CDFX. Individual patient-
tailored dosing of DFX should help to improve iron-chelation efficacy and to reduce dose-
dependent drug toxicity.
Key words. Deferasirox, bioavailability, thalassemia, iron chelation therapy, therapeutic drug
monitoring.
Abbreviations: BMI, Body mass index; DFX, Deferasirox; HPLC, high performance liquid
chromatography; LBM, lean body mass.
Introduction
The therapeutic benefit of iron chelation (i.e. forced elimination of iron) with deferoxamine (DFO)
has been clearly established in more than 40 years of clinical practice. For patients with thalassemia,
its introduction was quite literally life-saving, as this treatment has been shown to reduce iron-
related morbidity. However, DFO’s short plasma half-life (about 1 hr) and poor oral bioavailability
(less than 2%) necessitate slow subcutaneous or intravenous infusions with a pump. This procedure
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
is very demanding for patients and compliance is poor (less than 70%), even though compliance has
been shown to be absolutely crucial to obtain the full benefit offered by iron chelation (1, 2).
Research aimed at identifying new iron-chelating drugs with improved pharmacokinetic
characteristics and associated compliance, identified deferasirox (Exjade®, ICL670; Novartis Farma
S.p.A.), a once-daily oral iron chelator, which has been designed to treat chronic iron overload in
patients with transfusion-dependent anemias. Because of its convenient oral administration, DFX is
likely also to encourage compliance.
As reported by Waldmeier (3), deferasirox (DFX) showed biopharmaceutical, pharmacokinetic, and
metabolic properties that seem to be favourable for the intended purpose of chelation and
elimination of iron from the body; nevertheless, it can cause digestive disturbances (nausea, stomach
pain or diarrhoea) (4), or cause the patient to change the drug intake time to the evening (5).
Furthermore preparation of the solution from the dispersible tablets is particularly laborious. All
these factors can compromise patient adherence to therapy.
When DFX is administered orally to rats, at least 75% of the dose is absorbed but the bioavailability is
only 26% [3]. This difference between absorption and bioavailability after oral administration is
probably due to hepatic first-pass elimination (6).
The human pharmacokinetic parameters of DFX show a high intra- and inter-individual variability
that may affect the therapeutic response. In a clinical study in healthy volunteers, gastrointestinal
absorption was 70% (3); a median tmax value of 1–4 hours was observed in humans dosed with a
suspension formulation (3, 7). When taken at or close to meal-times, the type of food, its caloric
content and fat content (as well as the timing of the meal with respect to DFX administration) may all
influence its bioavailability and also increase the intra-subject variability of absorption. Therefore,
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
patients are recommended to administer DFX at least 30 minutes before a meal for optimum effect
(5).
The DFX and the iron complex are eliminated from the blood through the liver, largely in the bile.
With a mean elimination half-life (t1/2) of 8–16 hours, DFX plasma levels are maintained over 24
hours with once-daily administration (5, 7), but the inter-individual variability of t1/2 is large and
cannot always be estimated. Excretion of DFX and its metabolites occurs mainly within the first 24
hours and is complete within 7 days. The major DFX metabolic pathways are via direct
glucuronidation and conjugation of the hydroxylated metabolites with glucuronic acid and/or sulfate.
In the gut, the acyl glucuronide is unstable and undergoes hydrolysis, probably caused by microbial
glucuronidases, back to DFX, which leads to some enterohepatic recirculation (3). The presence of
large amounts of DFX in the intestine explains the observed reabsorption and enterohepatic
recirculation and the corresponding pharmacokinetics (5).
The DFX glucuronidation is under the control of UDP-glucuronosyltransferase 1A subfamily (UGT1A1
and, to a lesser extent, UGT1A3) (3, 6, 8); Gilbert Syndrome, a mild inherited form of
hyperbilirubinemia, that occurs in 15% of the population, is associated with a UGT1A1 gene
polymorphism. A homozygous insertion of TA pairs (genotype UGT1A1*28/*28) results in a decrease
in bilirubin glucuronidation activity and therefore leads to an increase in the level of unconjugated
bilirubin, although the precise role of the polymorphism is still not completely understood. The
UGT1A1 polymorphism has emerged as an important element in drug tolerance and could increase
the risk of toxicity of drugs metabolized via glucuronidation.
Dosing recommendations for most widely used drugs do not take into account adjustments for
individual characteristics. Knowledge of the gender and physical characteristics of a patient can be
essential for the optimisation of drug therapy, since these can substantially affect the drug’s
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
pharmacokinetic parameters and clinical effectiveness (9, 10). Limited preliminary results suggest the
possible influence of gender on the pharmacokinetics of DFX (11).
We report our preliminary experience on the potential usefulness of measurements of DFX plasma
concentration to optimize the drug’s dosage, on the basis of gender and physical characteristics;
individual patient-tailored dosing of DFX should help to improve the efficacy of iron chelation and to
reduce dose-dependent drug toxicity.
We have also evaluated the presence or absence of Gilbert Syndrome in patients, in order to assess
its effect on the efficiency of the glucuronidation metabolic pathway for DFX.
Patients and Methods
Patients
Patients of both sexes aged between 5 and 82 years with transfusion-dependent anemias
(thalassemia, myelodysplastic syndrome, or microdrepanocytosis), and who had been receiving daily
standard treatment with DFX for at least one year, were consecutively enrolled. Patients with any
neurological or psychiatric condition, with unstable or clinically significant cardiovascular disease,
with liver function tests (AST, ALT, bilirubin) ≥ 2 times the upper limit of normal (ULN), and with renal
impairment (creatinine clearance ≤ 30 ml/min according to Cockroft & Gault formula) were excluded
from the study. A signed informed consent was obtained before enrolment. The study was approved
by E.O. Ospedali Galliera Ethics Committee.
All patients received a single oral dose of DFX ranging from 8 to 42 mg/kg/day for at least three
months, and transfusion therapy once every 3–4 weeks. Liver, kidney and heart function and serum
ferritin concentrations were monitored every 4–6 weeks. At the screening visit, age, body weight,
height, and body mass index (BMI), were recorded and lean body mass (LBM) was estimated from
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
height and weight using the method of James (12). All patients were genotyped for the UGT1A1*28
gene polymorphism.
Only one blood sample per patient was collected in order not to change current clinical practice or
disrupt the patient’s lifestyle. Blood samples for measuring DFX plasma concentrations (CDFX) were
collected during a routine monthly visit, at the steady state, after the last DFX administration. For
each patient, an aliquot of 4 mL of blood was drawn into heparinized tubes; blood samples were
centrifuged at 1000g for 10 minutes and the resulting plasma was frozen and stored at -20°C until
analysis.
Chemicals and supplies
Deferasirox was kindly provided by Novartis Farma S.p.A., all reagents were of HPLC-grade and were
purchased from Merck (Darmstadt, Germany) and Sigma-Aldrich. The filtration system was obtained
from Millipore S.p.A. (Milano, Italy). A KromaSystem 2000 HPLC system consisting of a 325 pump
system, a 535 UV detector and signal integration software (BIO-TEK Instruments s.r.l. Milano Italy)
was used. Samples were analyzed on a 150 x 4.6 mm I.D. Alltima C18 column (Alltech Italia, s.r.l.,
Milano Italy) and a guard column (Alltech Italia, s.r.l., Milano Italy). Fresh human plasma samples
were obtained from healthy volunteers for standard samples.
DFX determination - Analytical procedure
Deferasirox plasma concentrations (CDFX) was determined with a validated HPLC assay previously
described by Rouan et al. Clinical samples, drug-free plasma and calibration standards were
extracted using this method (13). An aliquot of each extracted sample (100 µl) was injected onto the
HPLC column and eluted with a mobile phase (at pH 7) consisting of 0.05 M di-sodium hydrogen
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
phosphate and 0.01 M tetrabutylammonium hydrogen sulfate–acetonitrile–methanol (42:12:46,
v/v/v) at a flow rate of 1.3 ml/min at room temperature. The UV detector was set at 295 nm (ABS
0.1, RT 0.1); the run lasted for 15 minutes in total.
Calibration samples were prepared in pooled samples of blank human plasma and were prepared at
seven different concentrations ranging from 1.25 to 60 μg/ml. The results obtained from the analysis
of the calibration points were examined by linear regression. In order to assess whether a calibration
point could be accepted, it was back-calculated on the basis of the equation of the corresponding
calibration curve; a calibration curve was rejected if more than two concentrations or two adjacent
concentrations deviated more than 20% from the nominal value for the LLOQ (low limit of
quantification) and by more than 15% for the other concentrations (outliers). The calibration curves
of peak areas vs. concentrations of DFX were linear from 1.25 to 60 μg/ml, giving a correlation
coefficient r2 = 0.999. The results, as far as precision and accuracy are concerned, are derived from
the measured concentrations of the validation samples and were acceptable according to
Washington criteria (14).
Statistical analysis
The main endpoint variable considered was the level of CDFX. Summary descriptive statistics included
number (percentage) or rate of subjects for categorical data, mean ± standard deviation (SD) or
median and interquartile range (IQR) for continuous data. To visualize correlations between CDFX
(using the ratio CDFX/dose to adjust for DFX dose) and other factors we used scatterplots with fitted
linear regression lines; to test bivariate correlations, we calculated Spearman's rank correlation
coefficients and p for significance (figure 1 and figure 2). A linear regression model was built with CDFX
at response variable and age, gender, lean body mass, UGT1A1*28 gene polymorphism mutation
status, DFX dose, time from last drug intake to blood sampling and ferritin levels within 6 months
before as explanatory variable (table 2); interaction effects were tested among the factors
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
considered. A nonparametric K-sample test on the equality of medians and a nonparametric test for
trend across ordered groups (15) was used to compare CDFX over categories of UGT1A1*28 gene
polymorphism mutation status (figure 3). Two-tailed P value of 0.05 were adopted to define nominal
statistical significance; analyses has been conducted using STATA (version 13; StataCorp., College
Station, TX, USA).
Results
Eighty patients were consecutively enrolled and all successfully completed the study. Table 1
summarizes their demographic data, physical characteristics and laboratory test results. The median
total daily dose of DFX administered was 1500 mg, 25.8 mg/Kg/day (range 8-42 mg/kg/day), median
serum ferritin value in the 6 months before study initiation was 1100 ng/mL (interquartile range 717–
1916 ng/mL). With regard to presence of UGT1A1*28 gene polymorphism, in our cohort, about 40%
of all patients had the wild type allele, while 45% were heterozygous and 16% were homozygous.
Time from last drug intake to blood sampling followed a multimodal distribution, with a median
value of 3 hours (IQR: 1.4-14.0), with 3 main peaks at 2-4, 10-12 and 22-24 hours, as it is easily
deductible from figure 1, panel A. DFX median plasma concentration level was 17.1μg/mL (IQR: 8.8-
38.0): four patients (5%) were below LLOQ (equal to zero).
In figure 1 we show the scatterplots of CDFX/dose ratio vs. time from last drug intake to blood
sampling (panel A) and vs. age (panel B): both the variables were inversely associated with CDFX/dose
ratio (Spearman's rho=-0.25, p=0.03 and -0.18, p=0.1, respectively).
Concerning the association between plasma CDFX/dose ratio and ferritin blood concentration, and
between CDFX/dose ratio and lean body mass (LBM), scatter plots and fitted lines from univariate
linear regression are shown in Figure 2 (panel A and B, respectively). Both ferritin levels and LBM
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
appear to inversely correlate with DFX concentration levels, but the association was significant only
for ferritin (Spearman's rho=-0.36, p=0.002 and =-0.11, p=0.3, respectively).
No statistical correlation was detectable between CDFX and DFX dose (Spearman's rho =-0.06, p=0.6;
scatterplot not shown).
Multivariate linear regression model results are shown in table 2. The only two factors significantly,
independently and inversely associated with CDFX levels were time from last drug intake to blood
sampling and ferritin levels in the 6 months before study initiation (both p=0.006). There was a trend
maintain a consistent tense between age and CDFX, even if the statistical significance was not reached
when adjusted for the other factors (p=0.1). Other non significant factors in the model were LBM
(p=0.6), UGT1A1*28 gene polymorphism (p=0.9), gender (p=0.8), DFX dose (p=0.3).
To be noted, is the presence of an interaction between LBM (categorized as below and above the
median, 41.4 kg) and UGT1A1*28 gene polymorphism (three categories), which is significant (p=0.05)
in an univariate linear regression analysis, but remain close to the statistical significance (p=0.09)
adjusting for all other factors. We show this effect modification in figure 3, where a significant
inverse correlation (p=0.03) is evident between CDFX and UGT1A1*28 gene polymorphism, only in
patients with levels of LBM below the median.
Discussion
The goal of iron chelation treatment in patients with iron overload is to induce a negative iron
balance through the removal of excess iron deposited in organs. Reduction of total body iron is highly
dependent upon patient-specific factors including compliance, duration of iron overload, transfusion
burden and the specific properties of each chelator. There is still a need to find safe iron chelators,
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
which are able to remove iron from all tissues and which are available in formulations that ensure
maximum patient compliance. The human pharmacokinetic parameters of DFX suggest a high intra-
and inter-individual variability that may affect the therapeutic response and that could lead to
insufficient chelation or increased toxicity; some patients, particularly those with severe iron
overload, fail to respond adequately to DFX at the therapeutic doses, so various strategies have been
investigated to improve iron chelation in patients poorly responsive to DFX. In order to optimize the
drug’s efficacy and/or to monitor dose dependent adverse drug effects, it should be useful to
determine circulating levels of DFX.
Based on the present study results, DFX plasma concentration measured at different sampling times
showed a relatively high variability (interquartile range: 9-38 µg/mL) and is significantly associated
with the time from last drug intake to blood sampling (p=0.006, adjusting for all other factors
considered, table 2). Conversely, there was no linear correlation between CDFX and dose (p=0.3),
substantiating the prediction of a pharmacokinetic inter-individual variability and the observations of
several studies (11, 16). A question is whether no correlation mainly resulted from the inter-subject
pharmacokinetic variability, or also from poor adherence to treatment (i.e. incomplete dosing). Both
inter-subject pharmacokinetic variability and poor compliance have been widely demonstrated; in
our study other variables were taken into account that could affect the plasma concentrations of
DFX, such as gender and age; with regard to gender, in contrast to other authors (11), our results
seem to show a lack of influence of gender on the plasma concentrations, while we found a trend
toward an inverse association between CDFX and age (p=0.1). With regards to compliance, our
evaluation demonstrated DFX plasma concentrations below the quantification threshold (LLOQ) in
four patients (5%), although patients were being treated with therapeutic doses of DFX. All these
factors are a first step in demonstrating the need to monitor DFX plasma concentrations. The strong
inverse association between CDFX and ferritin concentrations (p=0.006) suggests that a higher plasma
DFX concentration could be associated with greater chelation efficacy (17). Since a correlation
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
between dose and DFX plasma concentration was not demonstrated, it seems useful to monitor the
concentrations to optimize and determine the most appropriate dose for each patient.
A further interesting result emerged from the analysis of genetic characteristics of patients. Evidence
in the literature suggests that a reduced transcription of the gene coding for UGT1A1, a situation
which is characteristic of Gilbert's Syndrome, may result in a reduction in enzyme activity of up to
70% in subjects with the allelic variant UGT1A1*28 (18-20). However, a mild reduction in the enzyme
does not always result in a full manifestation of Gilbert's Syndrome, which is, in any case, often
asymptomatic and neither diagnosed nor treated (19). Some recent studies have documented the
presence of other allelic variants of UGT associated with reduced enzyme activity (18-21). Although
these enzyme defects do not lead to clinically significant liver injury, they seem to play an important
role in regulating the metabolism of some drugs (18, 22, 23). It is therefore possible to hypothesize
that the presence of particular haplotypes of UGT could have profound effects on DFX disposition
and excretion.
In our study we tested the association between UGT1A1 polymorphisms and CDFX: we did not find any
significant association in the multivariate analysis (p=0.9, table 2), and, although a stratified analysis
of the correlation by mutation status was not feasible due to the low number of samples (lack of
statistical power), we found that LBM (categorized as below and above median value) was a
borderline significant effect modifier (p for interaction: 0.05, figure 3) of the relationship between
UGT1A1 polymorphisms and CDFX.
One hypothesis to justify this finding could be that, as a result of differences in percentage body fat,
lipophilic agents (i.e.DFX) may have a relatively greater volume of distribution (Vd), and water-
soluble compounds (i.e. glucuronate metabolite of DFX) a relatively lower Vd in subject with low
value of LBM. Therefore in subjects with low LBM the presence of UGT1A1 polymorphisms could
generate the lowest DFX plasma concentrations due to a higher distribution on body fat. The
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
expected analogous but inverse association (i.e. higher CDFX in UGT1A1 mutated patients with LBM
above the median value) was not observed in patients with LBM above the median value, probably
also because of the low statistical power.
Conclusions
Therapeutic drug monitoring (TDM) is an essential tool to optimize drug dosage; it is an aid to
rational therapy and has become essential in order to tailor the dosage to the individual. Thalassemic
patients might benefit from pharmacologic drug monitoring; measuring the plasma concentration of
DFX may be helpful because low concentrations could reflect either poor compliance or under-
treatment; individual patient-tailored dosing of DFX should help to improve iron-chelation efficacy
and to reduce dose-dependent drug toxicity. This study presented a simple, selective, sensitive, and
convenient method, potentially applicable to routine management of thalassemic patients; the
ruggedness of the analytical procedure ensures the validity and reproducibility of the assay during
standard clinical practice.The individual characteristics of patients could affect pharmacokinetic
parameters and clinical effectiveness of several drugs; for most widely used drugs, dosing
recommendations in adults do not take into account adjustment to individual characteristics. As
suggested by the European Medicines Agency (EMA) and the Food and Drug Administration Office of
Women's Health (FDA-OWH), which both emphasize the need to assess how demographic variables
influence the dose-response relationship defining aspects of the safety and efficacy of drugs (9, 10),
analyses of differences between the genders can provide useful information for possible use in
individualized therapy. Therefore the genetic and physical characteristics should be taken into
consideration during the optimization of drug therapy even if could be hardly feasible in the routine
clinical practice.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
Acknowledgements
The study did not provide any source of funding by the Sponsor.
Authorship Contributions:
Participated in research design: F Mattiolia, V Marinia, C Fucilea, L Robbianoa, S Perrottab, A Rosac,
Antonietta Martellia, G L Fornic
Conducted experiments: V Marini, C Fucile, G Milano
Performed data analysis: M Puntoni, F Mattioli, L Robbiano, A Rosa
Wrote or contributed to the writing of the manuscript: F Mattioli, M Puntoni, C. Fucile,
GL Forni
Conflict of interests: All authors declare that there are no conflicts of interest.
References
1. Olivieri NF, Brittenham GM. Iron-chelating therapy and the treatment of thalassemia. Blood
1997;89:739-61.
2. Meerpohl JJ, Antes G, Rücker G, et al. Deferasirox for managing iron overload in people with
thalassaemia. Cochrane Database Syst Rev 2012;2:CD007476.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
3. Waldmeier F, Bruin GJ, Glaenzel U, et al. Pharmacokinetics, metabolism, and disposition of
deferasirox in β-Thalassemic patients with transfusion-dependent iron overload who are at
pharmacokinetic steady state. Drug Metab Dispos 2010;38:808 –16.
4. Cappellini MD, Cohen A, Piga A, et al. A phase 3 study of deferasirox (ICL670), a once-daily oral iron
chelator, in patients with beta-thalassemia. Blood 2006;107:3455-62.
5. Galanello R, Piga A, Cappellini M D. Effect of food, type of food, and time of food intake on
deferasirox bioavailability: recommendations for an optimal deferasirox administration regimen. J
Clin Pharmacol 2008;48:428-35.
6. Bruin GJM, Faller T, Wiegand H, et al. Pharmacokinetics, distribution, metabolism, and excretion of
deferasirox and its iron complex in rats. Drug Metab Dispos 2008;36:2523–38.
7. Piga A, Galanello R, Forni GL, et al. Randomized phase II trial of deferasirox (Exjade, ICL670), a
once-daily, orally-administered iron chelator, in comparison to deferoxamine in thalassemia patients
with transfusional iron overload. Haematologica 2006;91:873-80.
8. Galanello R, Campus S, Origa R. Deferasirox: pharmacokinetics and clinical experience. Expert Opin
Drug Metab Toxicol 2012;8:123–34.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
9. European Medicines Agency (EMA) ICH –Gender considerations in the conduct of clinical trials
(EMEA/CHMP/3916/2005). London: EMA, 2005.
10. Obias-Manno D, Scott PE, Kaczmarczyk J, et al. The Food and Drug Administration Office of
Women's Health: impact of science on regulatory policy. J Womens Health (Larchmt) 2007;16:807–
17.
11. De Francia S, Massano D, Piccione FM, Pirro, et al. A new HPLC UV validated method for
therapeutic monitoring of deferasirox in thalassaemic patients. J Chromatogr B Analyt Technol
Biomed Life Sci 2012;893-894:127-33.
12. James WPT. Research on obesity (ISBN 0-11-4500347). London: Her Majesty’s Stationery Office,
1976.
13. Rouan MC, Marfil F, Mangoni P, et al. Determination of a new oral iron chelator, ICL670, and its
iron complex in plasma by high-performance liquid chromatography and ultraviolet detection. J
Chromatography B 2001;755:203–13.
14. Shah VP, Midha KK, Dighe S. Conference Report: Analytical Methods Validation: Bioavailability,
Bioequivalence, and Pharmacocinetic Studies. J Pharm Sci-US 1992;81:309-12.
15. Ref. Cuzick, J. A Wilcoxon-type test for trend. Statistics in Medicine 1985;4:87-90.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
16. Chauzit E, Bouchet S, Micheau M, et al. A method to measure deferasirox in plasma using HPLC
coupled with MS/MS detection and its potential application. Ther Drug Monit 2010;32:476-81.
17. Daar S, Pathare A, Nick H, et al. Reduction in labile plasma iron during treatment with deferasirox,
a once-daily oral iron chelator, in heavily iron-overloaded patients with beta-thalassaemia. Eur J
Haematol 2009;82:454-7.
18. Argikar UA, Iwuchukwu OF, Nagar S. Update on tools for evaluation of uridine
diphosphoglucuronosyltransferase polymorphisms. Expert Opin Drug Metab Toxicol 2008;4:879–94.
19. Bosma PJ, Chowdhury JR, Bakker C, et al. The genetic basis of the reduced expression of bilirubin
UDP-glucuronosyltransferase 1 in Gilbert’s Syndrome. N Engl J Med 1995;333:1171–5.
20. Mimura Y, Maruo Y, Ohta Y, et al. Effect of common exon variant (p.P364L) on drug
glucuronidation by the human UDP-glucuronosyltransferase 1 family. Basic Clin Pharmacol Toxicol
2011;109:486–93.
21. Maruo Y, Iwai M, Mori A, et al. Polymorphism of UDP-glucuronosyltransferase and drug
metabolism. Curr Drug Metab 2005;6:91–9.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
22. Lankisch TO, Behrens G, Ehmer U, et al. Gilbert’s syndrome and hyperbilirubinemia in protease
inhibitor therapy – An extended haplotype of genetic variants increases risk in indinavir treatment. J
Hepatol 2009;50:1010–8.
23. Ando Y, Ueoka H, Sugiyama T, et al. Polymorphisms of UDP-glucuronosyltransferase and
pharmacokinetics of irinotecan. Ther Drug Monit 2002;24:111-6.
Figures captions:
Figure 1. Scatterplots of CDFX/dose ratio vs. time from last drug intake to blood sampling (panel A) and vs.
age (panel B). To visualize correlations, parametric linear regression lines were drawn.
Figure 2. Scatterplots of CDFX/dose ratio vs. ferritin levels in the last 6 months (panel A) and vs. LBM values
(panel B). To visualize correlations, parametric linear regression lines were drawn.
Figure 3. Boxplot of CDFX and mutation for Gilbert’s syndrome, by LBM (median value).
Tables captions:
Table 1. Main characteristics of patients (n=80) undergoing measurement of plasma DFX
concentration.
Table 2. Multivariate linear regression model predicting CDFX with respect to patient characteristics.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
Table 1. Main characteristics of patients (n=80) undergoing measurement of plasma DFX concentration.
Age, yrs (mean ± SD) 31 ± 17
Sex, n (%)
Male 24 (30)
Female 56 (70)
Diagnosis
Beta thalassemia major or intermedia 71 (89)
Myelodysplastic syndrome 6 (7)
Microdrepanocytosis 3 (4)
UGT1A1*28 gene polymorphism (Gilbert Syndrome), n (%)
Wild type 31 (39)
Heterozygous 36 (45)
Homozygous 13 (16)
BMI, kg/m2 (mean ± SD) 22.5 ± 4.0
LBM, kg (mean ± SD) 40.7 ± 9.2
Ferritin levels 6 months before study initiation, ng/ml, (median, IQR) 1100 (717-1916)
DFX dose, mg/kg/day (median, IQR) 25.8 (20.0-32.6)
Time from last drug intake to blood sampling, hours (median, IQR) 3.0 (1.4-14.0)
DFX concentration (CDFX), µg/mL (median, IQR) 17.1 (8.8-38.0)
n: number of observations; SD: standard deviation; IQR: interquartile range (25th-75th percentile); BMI: body mass index; LBM: lean body mass; DFX: deferasirox.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
Figure 1. Scatterplots of CDFX/dose ratio vs. time from last drug intake to blood sampling (panel A) and vs. age
(panel B). To visualize correlations, parametric linear regression lines were drawn.
A
0
2
4
6
C DFX
/dos
e ra
tio
0 10 20 30Time from last drug intake to blood sampling (hours)
Spearman's rho=-0.25, p=0.03
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
B
Figure 2. Scatterplots of CDFX/dose ratio vs. ferritin levels in the last 6 months (panel A) and vs. LBM
values (panel B). To visualize correlations, parametric linear regression lines were drawn.
0
2
4
6 C D
FX/d
ose
ratio
0 20 40 60 80 Age (years)
Spearman's rho=-0.18, p=0.1
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
A
B
0
2
4
6
C DFX
/dos
e ra
tio
10 20 30 40 50 60 Lean Body Mass (Kg)
Spearman's rho=-0.11, p=0.3
0
2
4
6 C D
FX/d
ose
ratio
0 2000 4000 6000 8000 Ferritin levels 6 months before study initiation (ng/mL)
Spearman's rho=-0.36, p=0.002
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
Table 2. Multivariate linear regression model predicting CDFX with respect to patient characteristics.
Predictor β* 95%CI p
Age -0.2 -0.5 ÷ 0.03 0.1
Time from last drug intake to blood sampling
-0.7 -1.1 ÷ -0.2 0.006
Ferritin levels (6 months before) -0.005 -0.009 ÷ -0.002 0.006
Others not significant factors included in the model were: lean body mass (p=0.6), UGT1A1*28 gene polymorphism (p=0.9), gender (p=0.8), DFX dose (p=0.3). *β is the regression coefficients of standardized data coming from the model.
Acc
epte
d A
rtic
le
This article is protected by copyright. All rights reserved.
Figure 3. Boxplot of CDFX and mutation for Gilbert’s syndrome, by LBM (median value).
The bottom and top of the boxes are the first and third quartiles, the band inside the box is the median, and the ends of the whiskers represent the 9th and the 91st percentiles. Nonparametric K-sample test on the equality of medians of CDFX over UGT1A1*28 gene polymorphism categories: p=0.4 for LBM <41.4kg and p=0.8 for LBM ≥41.4kg.
Notes: *=test for the interaction between LBM* UGT1A1*28 gene polymorphism from the linear regression model on CDFX.
0
20
40
60
80
100
Wild-type
( 15)
Heterozygous
( 21)
Homozygous Wild-type
( 16)
Heterozygous
( 15)
Homozygous
( 9)UGT1A1*28 gene polymorphism (Gilbert Syndrome)
Lean Body Mass <41.4 kg (median) Lean Body Mass ≥41.4 kg (median)
C DFX
(µg/
mL)
P for interaction = 0.05*
P trend=0.03