`
`Safety profile of the oral iron chelator deferiprone:
`a multicentre study
`
`ALAN R. C oHEN .
`
`1
`
`3
`2
`4
`RE NZO G ALANELL0 ,
`ANTONIO PrcA,
`ANNUNZIATA DrPALMA,
`C ALOGERO V0110
`5 1Childre11's Hospital of Philadelphia, Philadelphia, PA. USA,
`A ND FER NANDO TRICTA
`2 Istituto di CU11ica e Bio/ogia Dell'Eta'Evolutiva. Cagliari. Italy, 3Universi ta Degli Studi di Torino, Torino,
`Italy, 4 Azienda Ospedaliera Arcispedale S. Anna', Ferrara, Italy, 5 Apo lex Research, Toronto, Canada
`
`4
`
`Heceivcd 12 May 1999; accepted for publication 14 October 1999
`
`Summary. In previous trials, the orally active iron chelator
`deferiprooe (Ll) bas been associated with sporadic agranulo(cid:173)
`cytosis, milder forms of oeut.ropenia aocl other side-effects. To
`detennioe the incideoce of these events, we perfo1111ed a
`multicentre prospective study of the chelator. Blood counts
`were performed weeldy, aod confirmed neutropeuia mandated
`discontinunrion of thera py. Among 187 patients with
`thalassaemia major, the incidence of agranulocylosis (neutro(cid:173)
`phils <0 ·5 x l09/ l) was 0 ·6/100 patient-years, aod the
`incidence of milder forms of neutropenia (neutrophils 0·5 -
`1·5x 109/l) was 5·41100 patient- years. All cases ofneutro(cid:173)
`penia resolved after interruption of therapy. Neutropenia
`occurred predominantly in non-splenectomized patients.
`Nausea and/or vomiting occurred early in therapy, was
`usually transient and caused discon tinuation of deferiprone
`in three patients. Mild to moderate joint pain an d/or swellin g
`
`did not require permanent cessation of deferiprone and
`occurred more commonly in patients with higher ferritin
`levels. Mean alanine transaminase (ALT) levels rose during
`therapy. Increased ALT levels were generally transient and
`occurred more com monly in patients with heparitis C.
`Persistent changes in immunological studies were infre(cid:173)
`quent., although sporadic abnorma lities occurred commonly.
`Meau zinc levels decreased during therapy. Ferritin levels did
`not change in the overall group but decreased in those
`patients with baseline levels> 2500 J.igl!. This study charac(cid:173)
`terized the safety profile of deferipron e, and, under the
`specific conditions of monitoring, demonstrated that agranu(cid:173)
`locytosis is less common than previously predicted.
`
`Keywords: tha lassaem ia , iron overload, chelation, deferi(cid:173)
`prone, safety.
`
`Iron overload is a predictable and life-threatening complica(cid:173)
`tion pf thalassaemia major and other haematological
`disorders whose successful management depends on the
`regular administration of red blood cell transfusions (Cohe1J,
`19 8 7). The iron chelator desferrioxamine was first intro(cid:173)
`duced more than 30 years ago (Sephton-Smith , 1962;
`;\.foeschlin & Schnider, 1963) and is still the only chelator
`approved for regular use in North America and. until recently,
`in Europe. Desferrioxamine reduces excessive iron, prevents
`iron-induced cardiac disease and improves survival in
`chronically transfused patients (Zurlo et al, 1989; Britten(cid:173)
`ham et al, 1994; Gabutti & Borgna-Pignatti, 1994; Olivieri
`et al, 1994; Olivieri & Britte1Jlrnm, 199 7). However, success(cid:173)
`ful therapy with desferrioxamine requ ires prolonged sub(cid:173)
`cutaneous or intravenous infusions of the chelator at least 4
`
`Correspondence: Alan R. Cohen, Chi ef, Division of Hematology. Tbe
`Children's Hospital of Phlladelpbia. 34th Street and Civic Center
`Boulevard. Phlladelphia, PA 19104. USA. e-ma il: cohen@email.
`cbop .edn.
`
`© 2000 Blackwell Science Ltd
`
`or 5 d each week (Gabutti & Piga, 1996), and compliance
`with this regimen has proven to be difficult for many patients
`as demonstrated by the continued development of iron(cid:173)
`induced cardiac dysrhythmias and heart failure (BritteDham
`et al, 1994; Olivieti et al, 1994; Gabutti & Piga, 1996). In
`addition, desferrioxamine may cause serious side-ellects, such
`as impaired vision, hearing loss, bone cha nges, growth retar(cid:173)
`dation and pulmonary toxicity (Olivieri et al, 1986: DeVirgiliis
`et al. 1988; Porter & Huelms, 1989; Freedman et al, 1990;
`Brill et al, 1991 ).
`After two decades of seeking an orally active chela ting
`agent as an alternative to desferrioxamioe, ooly one drug,
`deferiprone (l ,2-dimethyl-3-hydroxypyrid-4-one), has suc(cid:173)
`ceeded in entering extensive clinical trials. In rhese tria ls.
`deferiprone has decreased or stabilized the level of iron
`overload in most but not all patients with thalassaemia
`major (Kontoghiorghes et al, 1990; Toodury et al, 1990;
`Agarwal et al, 1992; Al-Refaie et al, 1992, 1995; Olivieri
`el al, 1995 , 1998; Hoffbrand et al, 1998). Adverse events
`
`305
`
`Coates
`Thursday, April 26, 2018
`
`Reported by: Elizabeth Borrelli
`CSR 7844, CCRR, CLR
`
`APO-FER-0203773
`
`1 of 8
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1053
`
`
`
`306 A . .R. Cohen et al
`associated with deferiprone in clinical trials have included
`agranulocytosis, arthropathy, gastrointestinal symptoms,
`increased ALT levels and progression of hepatic fibrosis
`(Hoffbrand et al, 1989: Bartlett ct al, 1990; Agarwal et. al,
`1992; Al-Refaie el al, 1992, 1994a, 1995; Berkovitch et al,
`1994; Olivieri el al, 1998). However, a clear salety prolile
`has not emerged from these studies because of the small
`number of patients, variations in study design and differ(cid:173)
`ences in drug formulation. To overcome these difficulties, we
`conducted a multiceutre study of a single formulation of
`defetiprone in patients with thalassaemia major. We designed
`the study to determine primarily the incidence of agranulo(cid:173)
`cytosis and other serious adverse events . As discussed below,
`the present study was not designed to address prospectively
`the concern regarding hepatic fibrosis that was raised after the
`completion of the trial (Olivieri et al, 1998). In this largest
`prospective study of iron chelation therapy to date, we
`also assessed the effectiveness of deferiprone. using serum
`ferritin levels to measure the SC\'erity of iron overload. A
`preliminary report has been published previously (Cohen
`et al, 1998).
`
`PATIENTS AND METHODS
`
`Patients
`Investigators in three thalassaemia centres in Italy (Cagliari.
`Ferrara. Torino) and one centre in the USA (Philadelphia)
`enrolled 18 7 patients with thalassaemia major in the study.
`The major elements of the inclusion criteria were: (1)
`transfusion-dependent thalassaemia; (2) age~ 10 years; (3)
`serum ferritin level > 2000 µg/l or liver iron concentration
`> 4 mg/g dry weight; and ( 4i an inability to use desferriox(cid:173)
`amine or an unwillingness to continue the use of the
`chelator despite medical advice. Exclusion criteria included:
`(1) a neutrophil count <2·0xl0 9 /l or a platelet count
`< lOOx 10 9/l during the past two years; (2) cardiac disease
`requiring medication; (3) hepatic or renal failure; (4) use of
`other iovestigational drugs and/or drugs known to cause
`neutropenia; (5) arthropathy; or (6) pregnancy or plans to
`become pregmmt during the study period. Patients with
`cirrhosis or chronic active hepatitis oo liver biopsy were also
`excluded, but liver biopsy was not a prerequisite for study
`entry except for the evaluation of hepatic iron concentration
`in patients with serum ferritin levels < 2000 µg/l.
`
`Study design
`The trial was a prospective open-label 1-year study of
`deferiproue. The primary objective was to detem1ine the
`incidence of agranu locytosis and other severe adverse
`events. The secondary objective was to determine the efficacy
`of deferiprone in the treatment of iron overload as assessed
`by serum ferritin level.
`Deferiprone was manufactured by Apotex (Toronto). the
`study sponsor, as 500-mg scored compressed film-coated
`tablets. Both the raw material and the finished dosage form
`were manufactured in Canada under Good Manufaclltring
`Practices (Gr.IP). A Data and Safety Monitoring Committee
`reviewed reports of serious adverse events such as agranu(cid:173)
`locyrosis and other cases of neutropenia.
`
`Deferiprone therapy
`The dose of deferiprone was 25 mg/kg body weight (bw).
`administered three times a day for a tota l daily dose of
`75 mg/kg bw. Each dose was rounded down to the nearest
`half tablet (250 mg) as required. The total daily dose was
`adjusted quarterly in accordance with chauges in body
`weight. Patients received a new supply of tablets every 4
`weeks, and they returned all unused tablets at the same time.
`Each centre maintained detailed drug accountability logs.
`Adhereuce Lo the treatment regimen was evaluated monthly
`as the percentage of prescribed tablets that were not
`returned.
`
`Definilions and moniloring
`A complete blood count (CBC) and white cell differential
`count were measured weekly (maximum 10-d interval). and
`each patient was questioned about the occurrence of any
`adverse events or the use of concurrent medications during
`the preceding week. If patients developed signs of infection,
`the CBC was determined twice weekly or more frequently as ·
`clinically indicated. If the neutrophils fell to < l ·5 x 10 9 /!,
`therapy with deferiprone was interrupted. and a CBC with a
`manual differential count was repeated the next day. ff the
`second neutrophil count was ~ l · 5 x 109 II. deferiprone
`therapy was reinstitu ted. If the two consecutive neutrophil
`counts were< l · 5 x 109 II. the patient was considered to have
`confinned neutropenia and deferiprone therapy was discon(cid:173)
`tinued. A confirmed neutrophil count < O· 5 x 109 /l was
`defiued as agranulocytosis.
`For patients with confirmed neutropenia. the neutrophil
`count was measured daily until it exceeded 2 ·0 x 109 II.
`Resolution of neutropenia was defined as two consecutive
`neutrophil counts> 2·0x109 /!, measured at least 3 d apart.
`At the time tl1e study was initiated, a platelet count
`< 100 x 109 /l was considered an adverse haematological
`event requiring discontinuation of therapy. After 3 months,
`t:he thresho ld was reduced to 50x109 /I because plcitelet
`counts of 50- lOOx 109/I were not uncommon in the study
`population and because this level was not considered to pose
`a significant risk.
`,
`Antinuclear antibody (ANA) titres. rheumatoid factors
`(RF), anti-double-stranded DNA (anti-dsDNA) and antihis(cid:173)
`tone antibodies (AHA) were obtained at baseline and every 3
`months. Serum ALT levels to assess liver function were
`obtained on a similar schedule. At baseline and every 6
`months, B lymphocytes. T lymphocytes, T-lymphocyte
`subsets and scrum zinc levels were measured. and patients
`undenvent a detailed rheumatological assessment. Serum
`ferritin levels to assess the degree of iron overload were
`measured at baseline and every 3 months. All baseline
`values were single measurements.
`
`Statistical analyses
`The statistical significance of changes from baseline for ALT
`level, B and T lymphocytes and T-lymphocytc subsets,
`plasma zinc couceutralion and serum ferritin level was
`detennined by a repeated measures analysis of variance,
`using the mixed model approach . The PROC MIXED program
`from SAS (SAS Institute. Cary. NC. USA) was used for these
`
`© 2000 Blackwell Science Ltd, British Journal ofHriematology 108: 305-312
`
`APO-FER-020377 4
`
`2 of 8
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1053
`
`
`
`'visit' as the repealed
`analyses. The model contained
`measure and 'patient' as the random factor. The covariance
`structure was autoregressive of order 1. If 'visit' was shown
`to be statistically significant (P < O·O 5 ), the CO NTRAST state(cid:173)
`ments comparing the means from various visits to that at
`baseline were examined to determine the significance of
`change from baseline. For ALT and serum ferritin levels. the
`analyses were conducted on the data set for all patients
`(intention-to-treat · analysis) as well as the data set for
`completers only. In the intention-lo-treat: analyses, the last
`available data for patients who 1"7ithdrew during the study
`were carried forward to the subsequent time points. For the
`analysis of zinc conceDtratioll and lymphocyte number, only
`the data set for complelers was used.
`The change in incidence of ALT levels greater than twice
`the upper limit of the reference range was analysed by
`logistic regression analysis. The difference in incidence of
`ALT levels greater than twice the upper limit of the reference
`range in patients who were seropositive and seronegative for
`hepatitis C was analysed by the chi-squared test. The
`significance of the association between arthropathy and
`serum ferritin level was determined by the Mantel-Haenszel
`chi-squared test. The trend of serum ferritin level in each
`patient was determined by simple linear regression. The
`Fisher's exact test was used to evaluate the difference in
`incidence of neutropenia between splenectorn.ized and non(cid:173)
`splenectomized patients. The significance of the differences
`in the neutrophil .count between splenectomized and non(cid:173)
`splenectomized patients at baseline and during the study was
`determined by analysis of covmiance. using the mixed model
`approach. The model contained ·spleen status' as the fixed
`factor and 'patient' as the random factor. The covariance
`structure used was autoregressive of order 1.
`All statistical tests were two-sided with a type I error {a) of
`0 ·05.
`
`Informed consent.
`The study, designed in accordance with the staudards of
`Good Clinical Practice, was approved by the institutional
`review board of e;:ich of the participating centres. lnformed
`coasent was obtained from all patients and/or
`their
`parents.
`
`RESULTS
`
`Study patients
`One hundred and eighty-seven patients with thalassaemia
`major, ranging in age from 10 to 41 years (mean 18 ·4
`years), were enrolled in the trial. Seventy-four patients (40%)
`h;:id previously undergone splenectomy, and 142 (76%) were
`seropositive for hepatitis C. Regular chelation therapy with
`desfenioxamine was not presctibed for six patients at the
`time of study entry. The mean prescribed dose of desferriox(cid:173)
`amine in the remaining 181 patients was 40 mg/kg bw, with
`a mean frequency of 6· 5 prescribed infusions per week.
`
`Drug exposure
`One hundred and sixty-two patients completed the 1-year
`study. Two patients withdrew, two were removed because of
`
`307
`Safety Study of Deferiprone
`protocol violations, and 21 stopped therapy after the
`occurrence of adverse events described previously (Cohen
`et al, 1998). The mean number of prescribed days of
`defetiprone for the entire study population was 325. yielding
`a cumulative total of 168 patient- years of drug exposure.
`The mean total daily dose was 73 mg/kg bw. Adherence to
`therapy, as assessed by pill counts, was 93 = 8% (range
`40- 100%).
`
`Urinary changes
`Reddish discoloration of the mine, the most cotmnonly
`reported event (74 patients), began as early as the first week
`of treatment. The finding persisted or recurred sporadically
`throughout the study period. The chauge in urine colour
`was not accompanied by dysuria, increased frequency or
`other urinary symptoms.
`
`Gastrointestinal symptoms
`Nausea and/or vomiting (45 patients) occurred most
`commonly during the first few weeks of therapy, were
`usually transient, were mild to moderate in severity and
`generally required no alteration in treatment. However.
`nausea caused interrupt.ion of therapy in four patients (2%)
`and subsequent discontinuation in three. Abdominal pain
`(26 patients) was mild to moderate in severity and resolved
`without discontinuation of therapy.
`
`A rtllropathy
`Twenly patients reported joint pain, two reported joint
`swelling and two reported both symptoms. In 14 of these
`patients, the complaints were judged to be possibly related to
`the study drug. The onset of joint symptoms varied from 7 to
`3 61 days after study entry, and all episodes were considen,'C!
`mild or moderate in severity. The median time to resolution
`was 12 days. Joint symptoms persisted for 2'. 3 weeks in five
`patients. Therapy with deferiprone was discontinued tem(cid:173)
`pornrily iu four pal.ienl.s imd I.he symptoms resolved wit.bin
`13 clays. All four patieuts restarted therapy with deferipron e;
`in one patient, mild arthralgia recurred but required no
`change in
`treatment. None of these 24 patients had
`abnormal joint lindings on rheumatological assessment at
`baseline, at 6 months, or at 12 months. The frequency of
`reported arthropathy increased with the level of iron
`overload (P = 0·03). Arthropathy was reported in 9% of
`patients with baseline ferritiu levels less than 2500 µgll.
`15% of patients with baseline ferritin levels of 2 500-
`5000 ;ig/l, and 26% of patients with baseline ferritin levels
`greater than 5000 µgll.
`
`Alanine transaminase (ALT) levels
`ALT values in individual patients £luctuated throughout the
`trial. with declines from baseline of up to 369 U/l and
`increases from baseline of up to 5 64 U/l. For patients who
`completed the study, the mean ALT was significantly higher
`than
`the baseline level of 59 :!:: 65 U/l at 3 months
`(76=72U/I; P=0·007) and at 6 months (78±:81U/I;
`P = 0·002). For all patients (intention-to-treat analysis),
`the mean ALT was significantly higher than the baseline
`level of 61 :!:: 63 U/I at 3 months (81::: 73 U/I; P= 0·0001).
`
`© 2000 Blackwell Science Ltd. Britislz Journal ofHaematology 108: 305-312
`
`APO-FER-0203775
`
`3 of 8
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1053
`
`
`
`308
`
`ii.. R. Cohen et al
`Table I. Number of patients witb ALT levels> 2x upper limit of nonnal. according to hepatitis C serostatns.
`
`Hepatrns C serostatus
`
`Baseline
`
`3 months
`
`6 months
`
`9 mon\bs
`
`Termination
`
`Negative
`Positive
`Total
`
`4
`39
`43
`
`3
`43
`46
`
`6
`38
`44
`
`6
`41
`47
`
`6
`4~
`49
`
`Table II. Number of patients with posifjve tests ror antinuclear antibody (ANA). anti-double-stranded DNA (anti-dsDNA), rhenmatoid factor (RF)
`and antihistone antibody (AHA).
`
`Positive at baseline and
`negative ai termination
`
`Positive at baseline
`and at termination
`
`New and transient
`positive
`
`New and persistent
`positive
`
`New positive at
`completion only
`
`ANA
`Anti-dsDNA
`RF
`AHA
`
`6
`[)
`
`[)
`1
`12
`1
`
`2
`
`26
`2
`
`n
`3
`3
`[)
`
`2
`6
`
`(8 1 ::!: 79 U/l; P= 0·0005) and 9 months
`6 months
`(76::!:63U/l; P=0·014). The proportion of patients with
`ALT levels greater than twice the upper limit of the reference
`range did not differ at each quarterly assessment (P = 0·4 3;
`Table I). ALT levels greater than twice the upper limit of the
`reference range were present in 43 patients (23%) at
`baseline and in S 5 additional patients (29%) on at least
`one measuremeut during the study. Eighty-three of these 98
`patients (85%) were seropositive for hepatitis C, compared
`with 59 of the 89 patients (66%) with ALT levels less than
`twice the upper limit of the reference range (P = 0·003 ).
`Four patients. all seropositive for hepatitis C. interrupted
`therapy with deferiproue as a result of increased ALT levels.
`The ALT levels subsequently returned to baseliJJe values. 0De
`of the four patients discontiJJued therapy without rechal(cid:173)
`leDge. llpoD reintroduction of deferiproDe in the other three
`patients, ALT levels rose again and treatment was discon(cid:173)
`tinued in one patient.
`
`Agrariulocylosis and olher rie11Lropenias
`Agranulocytosis (neutrophils < 0· 5 x 109 /I) occurred in 1
`out of 187 patients (0·5%). The overall incidence was 0·6/
`100 patieJJt- years of therapy. Less severe forms of confirmed
`
`neutropenia (neutrophils 0·5 - 1·5 x 109 /1) occurred in nine
`patients (4·8%). giving an incideJJce of 5·41100 patient(cid:173)
`ycars of treatment. All cases of agranulocytosis and
`JJeutropenia resolved after discontiJJuatioJJ of deferiproJJe.
`The clinical courses of affected patients have been described
`previously (CoheJJ et al, 1998).
`The incidence of neutropenia in splenectomized patients
`was l ·6/100 patient-years and the incidence in non-splenec(cid:173)
`tomized patients was 8· 71100 patient-years (P = 0·09) . The
`mean neutrophil count at baseline in Don-splenectomized
`patients (4·0::!:l·5x109/1) was significantly lower than iD
`spleneclomized patients (6· 3 ::!: 3 · l x 109 / 1: P= 0·0001) and
`remained so throughout the study (Fig 1).
`
`Immwwlogical studies
`The number ofB lymphocytes and CD3+, CD4+ and CDS +
`T lymphocytes. measured by flow cytometry at 6 months and
`at the completion of the study period. did not differ from the
`baseline values (P>0·05). Twenty-four patients (12·8%)
`had positive results for ANA. anti-dsDNA, RF or AHA at
`baseline (Table II). and results remained positive in 14 palients
`at completion of the study. Transiently positive results during
`therapy with deferiproDe were common. occurring in 33
`
`8
`
`7
`
`"' 0
`6 R
`<: 5
`::l
`0 4
`(..)
`£ 3
`0. _g 2
`Q) z 1 e
`0 ...
`0
`
`::l
`
`12
`
`24
`Time (weeks)
`
`36
`
`48
`
`Fig 1. Mean weekly neutropbil count in 74
`splenectomized and l l 3 non-splenectomized
`patients. The mean neutrophil count in non(cid:173)
`splenectomized patients was significantly
`lower tlian in splenectomized patients
`(P=D·DDDl) from study entry to completion.
`
`© 2000 Blackwell Science Ltd. British Joumal of Haematology 108: 305-312
`
`APO-FER-0203776
`
`4 of 8
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1053
`
`
`
`8000
`
`7000
`
`6000 .
`
`c:
`
`~ 5000
`"" -~ 4000
`u.
`E 3000
`2
`Q) 2000 .'
`(/)
`
`1000
`
`0 ·'
`
`>5000
`T
`
`Safety Study of Dejeriprone
`
`309
`
`$2500
`
`2501 -5000
`
`I
`
`Baseline 12 Months Baseline 12 Months Baseline 12 Months
`
`•••••••• •
`
`Fig 2. Mean sernm ferritin le1,els at baseline
`and at 12 months according to ferritin levels at
`study entry.
`
`patients (17·6%). Six patients (3 ·2 %) who had negative
`studies for ANA, anl i-dsDNA, RI' or AHA at baseline sub(cid:173)
`sequently developed a new antibody that persisted on the
`next measurement. In addition, ten patients (5·3%) had a
`first positive antibody at the final assessment.
`
`Zinc levels
`The plasma zinc levels at baseline (14·4 = 2·3 µmoll!)
`decreased to 13 ·9 ::'::: 2·3 µmol/l at 6 months (P= 0·013)
`and to 13·0:::2·lµrnol/l at 12 months (P =O·OOOl).
`Thirteen of the 171 patients with normal levels at baseline
`had values below the reference range at 6 months only, 24 at
`study completion only, and seven at both 6 months and
`study completion.
`
`Efficacy
`In the l 62 patiems who completed l year of therapy, the
`mean serwn ferrilia levels at baseline and at 12 months
`were 2579 ::'::: 1777 µg/l and 2452 ::'::: 1452 itg/l respectively
`(P = O· 2 6 ). The values at baseline and termination for all
`patients who received deferiproue, whether or not they
`completed the study (intention-to-treat analysis). were
`2696::':::1877 µg/l
`and 2633::':::1815µg/l
`respectively
`(P=0·58).
`Baseline serum ferritin levels were stratified using a pre(cid:173)
`viously defined threshold for development of cardiac disease
`(Olivieri et al, 1994). For patients with baseline ferritin levels
`,;; 2500 µg/1 who completed 1 year of therapy (n = 99), the
`mean ferritiu levels at baseline and at 12 mouths did not
`differ (P = 0 ·2 8; Fig 2). For patients with baseline ferritin
`levels> 2500 µg /l who completed l year of therapy (n = 63),
`the mean ferritiu level declined from baseline to 12 months
`(P = 0·001). A subanalysis of patients with baseline ferritin
`levels> 5000 µg!I (n = 15) demonstrated a decline in ferritiu
`level fmm 7042 ::'::: 2143 µg/l at baseline to 5094 ::'::: 1334 µg/I
`ar 12 months (P=0·0008). Simple Linear regression indi(cid:173)
`cated a negative slope in 5 7% or the 162 patients who
`completed the study.
`Therapy 1~ith deferiprone was interrupted temporarily
`in six parieuts when their ferritin levels fell below 500 p.g/l
`and was reinstituted when the ferritin level exceeded this
`threshold. The mean time of interruption was 62 ::: 20 d.
`
`DISCUSSION
`
`to characterize
`This multicentre study was designed
`prospectively the incidence or adverse events during therapy
`with deferiproue and the relationship of these events to the
`ch el a tor. In particular, enrolment of 18 7 patients · and the
`weekly monitoring of blood counts enabled this study to
`determine, for the first time. the incidence of agrauulo(cid:173)
`cytosis. the most serious side-effect associated with deferi(cid:173)
`prone. This 1-year trial, completed hefore the question or
`progression of hepatic fibrosis during therapy with deferi(cid:173)
`prone arose (Olivieri el al, 1998). did not require liver
`biopsies and therefore cannot address this issue. Additional
`studies to answer this question are presently u1Jder way.
`The most frequently reported adverse event in this study
`was a reddish discoloration of the urine. The colour was
`atrributed to the excrerion of the iron - deferiprone complex.
`which is a chromophore (K ontoghiorghes et al, 198 7). No
`patient discontinued therapy with <leferiprone because of the
`change in urine colour. Gastrointestinal symptoms, the
`second most commonly reported adverse event, usually
`occurred during the first weeks or therapy and resolved
`without altering the dose of deferiprone. These findings are
`similar to those reported by other investigators (Al-Rcfaie
`et al. 1992, 1995).
`lo previous studies, the frequency of arthralgia and
`arthritis during therapy with deferiprone has varied from
`14% to 38% (Agarwal et al, 1992; Olivieri et al, 1995). The
`International Study Group on Oral Iron Chelators (ISGO!C)
`found joint pain or stiffness in 20% of patients in their
`com bined data pool (Al-Refaie el al, 1995). ln these studies,
`kuee pain was the most common complaint, sometimes
`accompanied by joint swelling and loss of motion. Magnetic
`resonance imaging and synovial biopsies have not estab(cid:173)
`lished a pathological basis for the joint abnormalities, and
`serological stu dies have failed to detect evidence of auto(cid:173)
`immune disease in most: patie1Jts (Berkovitch et al, 1994:
`Hoflbrand el al, 1998). In the present study, the frequency or
`arthropathy (13%) was less than previously reported. Four
`patients temporarily interrupted therapy because of joint
`pain. hut no patient discontinued trearment for this reaso1J.
`A previous report suggested that deferiprone-associated
`arthropathy is more frequent in patients who are more
`
`© 2000 Blackwell Science Ltd. Britisli Journal of Haematology 108: 305-312
`
`APO-FER-0203777
`
`5 of 8
`
`Taro Pharmaceuticals, Ltd.
`Exhibit 1053
`
`
`
`310 A. R. Cohen et al
`heavily overloaded with iron (Berkovitch el al, 1994). The
`present study supports this theory by finding an increased
`frequency of joint abnormalities at higher ferritin levels.
`Three cases of autoimmune disease during treatment with
`deferiprone have been reported in patients with thalassaemia
`(Mehta el al, 1991; Castriota-Scamlerbeg & Sacco, 199 7). In
`the present study, new and persistently positive tests for ANA,
`anti-dsDNA, AHA and RF were uncmrunon and were not
`associated with other findings suggestive of autoimmune
`disease. These results are consistent with earlier clinical trials
`that forn1d no clinical or laboratory evidence of deferiprone(cid:173)
`induced systemic lupus erythematosus (Bartlett et al, 1990;
`Berdoukas, 1991; Al-Refaie et al, J 992: Berkovitch et al,
`1994: Olivieri et al. 1995).
`Low plasma zinc levels have been reported in some
`patients receiving deferiprone, particularly those with diabetes
`mellitus (Al-Refaie cl al. 1994b, 1995; Hollbrand et al, 1998).
`ln the present study, despite the absence of zinc supplemen(cid:173)
`tation, most patients maintained nonnal zinc levels, but the
`mean level for the group declined. Seven patients (3·7%),
`including one of three patients with diabetes, developed low
`plasma zin.c levels at 6 months that persisted at. 12 months.
`These findings indicated that low plasma zinc levels were not
`a common problem in patients with thalassaemia who were
`treated with deferiprone, but periodic monitoring of plasma
`zinc levels, especially in patients with diabetes, may be
`warranted.
`Scrum transaminase levels increase transiently during
`therapy with deleriprone in some patients (Al-Refaie et al,
`1992, 1995; Olivieri et al, 1995). The lSGOIC reported ALT
`levels at least t"'ice the upper limit of the reference range in
`60% of patients (Al-Refaie et al, 199 5) . This increase
`occurred most frequently during the first 6 months of therapy,
`was mild and resolved without alteration of treatment in the
`majotity of patients. In the present trial, ALT levels at baseline
`exceeded twice the upper norm al limit in 2 3 % of patients. and
`new eleva1.io11s of ALT during therapy with deleriprone
`occurred in 29% of patients. Most of these episodes were
`transient and occurred in patients who were seropositive for
`hepatitis C. The clinicnl significance of this effect of deferiprone
`on the serum ALT levels in some patients remains unclear.
`Olivieri et al (1998) have reported progression of fibrosis in
`patients receiving deferiprone (Olivie1i et al, 1998), and have
`cited studies in gerbils with another hydroxypyridone
`(Carthew el al. 1994) and studjes in l1lro of deferiprone
`(Cragg et al, 1998) in support of this relationship. Acknowl(cid:173)
`edging the need for additional studies, some investigators
`bave questioned tbese conclusions (Kowdley & Kaplan,
`1998), and others have expressed reservations about the
`usefulness of the gerbil model (Ramm et al, 1999). Still other
`investigators have found no association between deferiprone
`and hepatic fibrosis in patients receiving the chelator (Callea,
`1998; I-loffbrand et al, 1998; Piga el al, 1998; Stella et al,
`1998; Tondury et al. 1998).
`The most serious adverse effect of therapy v.ith deferiprone
`is agranulocytosis, with a previously estimated incidence of
`l ·8/100 patient- years (Al-Refaie et al, 1995). The estimated
`incidence of less severe forms of neutropenia is 2 ·41100
`patient- years (Al-Refaie et al, 1995). The mechanism of
`
`deferiprone-associated agranulocytosis re1m1ins uncertain .
`Jn animal studies, concomitant falls in the white cell count,
`red cell count, haemoglobin level and platelet count suggested
`a generalized myelosuppressive effect of the chelator (Konto(cid:173)
`ghiorghes et al, 1989; Port.er el al, 1991; Berdoukas cl al,
`1993). In non-iron-loaded rats and dogs , the myelosuppres(cid:173)
`sive effect of deferiprone appears to be dose-dependent
`(Kontoghiorghes et al. 1989; Biesemeier & Laveglia, 1991).
`No evidence for an immune mechanism has been demon(cid:173)
`strated \Al-Rel'aie el al, 1993, 1994a), and no increase in the
`sensitivity of myeloid precursors to deferiprone in vitro was
`observed in the investigation of a patient who developed
`agrmrnlocytosis during therapy with the chelator (Cunning(cid:173)
`ham el al. 1994). Nevertheless. in several reports, rechalleuge
`with deferiprone of patients who developed agranulocytosis
`or milder neutropenias has been associated with recurrent
`changes in the neutrophil count (Al-Rel'aie et al, 1994a;
`Hollbrand. 1996; Olivieri. 1996).
`In the present study, which used a single preparation of
`deferiprone manufactured according to GMP standards, a
`uniform dose and weekly monitoring of blood counts, the
`incidence rates of agranulocytosis and milder forms ofneutro(cid:173)
`penia were 0·6/J 00 and 5·4/J 00 patient- years respectively.
`Some evidence from this trial suggests that the presence or
`absence of the spleen may influence the occurrence of neutro(cid:173)
`penia. Although this evidence must be interpreted with
`caution because of the small number of patients and the
`absence ofrechallenge with deferiprone. the spleen may play
`a role in determining the likelihood of neutropenia. In non(cid:173)
`splenectomized patients whose neutrophil counts are lower
`than in splenectomized patients. normal variation in white
`cell count and factors such as viral illnesses may increase the
`probability of a fall in neutrophil count to a level defined as
`neutropenia. Weekly haematological assessments may detect
`more cases of neutropenia by uncove1ing transient falls in
`the neutrophil count that would be missed with less frequent
`monitoring. Nevert:hele~..,;. in lhis study, the frequent: labora(cid:173)
`tory testing and the permanent cessation of dcfcriprone
`therapy after detection of a confinned neun·ophil count below
`] · 5 x] 0 9 /I may have prevented more severe forms of neutro(cid:173)
`penia and may have led to an underestimation of the
`incidence of agranulocytosis that might occur with less
`frequent monitoring and with a different threshold for
`discontinuing therapy.
`The secondary outcome of t