i•I;IICJ0Mif4Qiil
`
`Clinical and Experimental Allergy, 36, 676- 684
`
`© 2006 Blackwell Publishing ltd
`
`Effects of levocetirizine as add-on therapy to fluticasone in seasonal allergic
`rhinitis
`M. L Barnes, J. H. Ward, T. C. Fardon and B. J. Lipworth
`A5thma Et Allergy Research Group, Deportment of Medicine and Therapeutics, Ninewelfs llospital Et Medico/ School, Univet5ity of Dundee, Dundee, Scot/and, UK
`
`Clinical and
`Experimental
`Allergy
`
`Corrtspondence:
`Mr Martyn L. Barnes, Asthma & Allergy
`Research Group, Departmen of
`Medicine and Therapeutics, Ninewells
`Hospita l & Medical Sch ool, University of
`Dundee, Dundee 001 9SY, Scotia nd, UK.
`E-mail : mbarnes@ rcsed.a ~. uk
`
`Summary
`Background Addition of H 1 antagonists to intranasal corticosteroid treatment of allergic
`rhinitis (AR) is common in
`linical pra tice and recommended by guidelin s, despite som
`vid n "l' that th ' additiv ben f1ts ar n gligibl .
`Objective To assess additional benefits of 5 mg levocetirizine dihydrochloride in seasonal AR
`pati ··nt using 200 meg fluti a on propionat nasal pray one , daily.
`Method In a douhk-blind pia · bo-controll d crossover study of 27 patients, fo ll ow ing 2
`weeks without treatment, subjects used flulicasone with levocetirizine or identical plac bo for
`2 weeks each. Assessments were the Juniper mini Rhinoconjunctivitis Quality-of-Life
`Questionnaire (min.i-RQLQ), domiciliary peak nasal inspiratory flow (PNLF), total nasal
`symptoms (TNS) scores and nasal nitric oxide concentratiollS. Effects were interpreted and
`t st d aga inst minimal clinically important differ nces.
`Results Add-on effects for levocetirizine vs. placebo excluded any clinically significant
`bem•f1ts: m an ffects (on sid d 950fo conf1d n
`intervals) were mini-RQLQ - O. ll (- 0.34),
`PNTF +0.57 ( + 5.23), and TNS - O.ll (- 0.60). Numbers needed to tr at (950/o conf1dence
`int rvals) by ou tcom e were min i-RQLQ 14 (5 to 49), PN IF 4 (3-7), and TNS 3 (2-6). No
`sign.iJJcant within or between treatment fti cts were seen for nasal nitric oxide.
`Conclusion Contrary to current practice, the present. results demonstrate thai for the majority
`of paLlents, antihistamine add -on to effective nasal steroid t:reatmcnl is inappropriate. Further
`work is required to confum that this is also true in the most severe cases, and the available
`evidence needs to lJe pul into guidelines and implemented.
`
`Keywords clin ical rel evance, fluticasone propionate. levocetirizine dihydroch loride, minimal
`clinically important difference, seasonal allergic rhinitis
`Submitted 19 Augusc 2005; revised 6 .January 2006; accepted 27 .January 2006
`
`Introduction
`
`Both antihistamines and nasal steroids are commonly
`used in the treatment of allergic rhinitis (AR). There is
`good evidence of efficacy for monotherapy with either,
`but the response is frequently incomplete and patient
`sati fa tio n is poo r, o combination th rapy i oft n
`start d. The ARIA gu id lines [I] comm nt 'The tr atm nt
`of all rgi rhiniti
`impli s symptom reduction by drug
`and att mpts to interf r in the intlanunatory cascade by
`
`The study was funded by a departmental grant from the asthma
`and allergy resea rch group and received no fi nancial support from
`the pharmaceutica l industry.
`
`anti- inflammatory drugs or specific immunotherapy ...
`theoretically, combining imerventions at differenr levels
`should improve the clinical outcome'. The same report
`described differential effects for intranasal corticosteroids
`and antihistamin s on symptom , so it would s em
`intuitive that additive effect could exist.
`and
`A lit ratu r
`·arch (M dl in , Co hran , Emba
`anc stor referenc s) wa conduct d for evidence upport(cid:173)
`ing add-on of an ihjstamin s fo r patien using ropical
`na al t roid
`in AR.
`In a recent review, Akerlund et a l. [2) hjghlighted the
`lack of evidence to support newer guidelines that recom(cid:173)
`mend the add-on of antihistamine to nasal steroid
`therapy.
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`PLAINTIFFS'
`TRIAL EXHIBIT
`PTX0012
`
`MEDA APTX02173909
`
`PTX0012-00001
`
`CIPLA LTD. EXHIBIT 2024 PAGE 1
`
`

`

`A r view by Howarth [3] obs rv d that th re was no
`evidence from 'limited studies available' to support super(cid:173)
`iority of combination therapy compared with topical
`corticosteroid alone.
`Ratner
`t al. [4] evaluated 2 week treatments with
`loratadine ( 10 mg once daily) and nasal fluticasone
`(200 meg once dany) alone or in combination in moderate
`to severe seasonal AR and found no evidence of 'mean(cid:173)
`ingful' additional ben fJtS on ymptoms co r sand qual(cid:173)
`ity of life.
`Di Lorenzo et al. [51 compared the use of nasal flutica(cid:173)
`sone (200 meg once daily) alone and in combination with
`cetirizine (10 mg once daily) in moderate to severe seaso(cid:173)
`nal AR, showing that combination therapy r suit d in a
`statistkally significant but small improvement in nasal
`itching and combined symptom score . No significant
`improvem nts wer se n for ong stion, rhinorrhoea,
`sneezing, or percentage of eosinophils and eosinophil
`ationic protein in nasal washings.
`Notably, none of the studies identified in our search
`conducted equivalence or non-superiority analys s com(cid:173)
`paring effects to defined limits of clinical relevance, and
`so were not able to conclude that combination therapy is
`clinically inappropriate.
`ertajnly mo r xp n iv , and
`Combination therapy i
`inh rent risks of polypharmacy, including poor
`has th
`ompliance, jnt ' raction
`;md addition al sid -effe ts,
`which although rare and usually mild, do include seda(cid:173)
`tion, palpitations, arrhythmias and hypersensitivity reac(cid:173)
`tions.
`We therefore conducted a double-blind placebo-con(cid:173)
`trolled crossover study of the effects of fluticasone pro(cid:173)
`pionate alone or in combination with levocetirizine.
`Outcome assessments included bot.h objective and sub(cid:173)
`jective measures. The primary outcome was the Juniper
`mini-RQLQ [6]. We conducted a non- -up ·riority analysi
`for levocetirizine vs. placebo as add-on to fluticason e
`to minimal cl inically important
`treatment, with refcrcnc
`differences (MCIDs} for each outcome.
`
`Methods
`
`Patients
`
`Participants were identified from our own database of
`patients in the Dundee area, Scotl and. Inclusion criteria
`were male or female patients, aged 16-75 yea.rs, with
`seasonal (intermittent or persi t nt) AR, and kin prick(cid:173)
`po itive responses to grass pollen. Exclusion criteria were
`any other conditions affecting nasal airway patency,
`including septal deviation greater than 500/o, and grade
`2 polyps (extending below the upper cdg of the inferior
`turbinate), pregnancy, lactation or any medical condition
`or screening blood result that might compromise partici(cid:173)
`pant safety.
`
`© 2006 Blackwell Publishing Ltd, Clinical and Experimental AJ/ergy, 36 : 676- 68 4
`
`levocetirizinc as add-on -hcrapy to fluticasone in AR 6 7 7
`
`Run-iu
`(2 weeks)
`
`T1ealmtnl J
`(2 weeks)
`
`T'rearl'm:nt 2
`(2 weeks)
`
`Treatments, all taken m the morning:
`rP (Fiuticasone nasal spray) - two sprays to each nostril (200 meg totaQ.
`LEVO (Levocetirizlne tablets) - 5 mg.
`
`Fig. 1. Study Oow diagram.
`
`Study design
`
`The study (se Fig. l) was conducted in four vi its to the
`research laboratory in the months of June and July 2004-
`the Dundee peak grass pollen season. All participants gave
`informed consent, and the protocol was given a favour(cid:173)
`able opinion by the multi-centre research ethic commit(cid:173)
`for Scotland.
`t
`Vi it on (scr ning) d t rmin d in lu ion and exclu(cid:173)
`sion status. A m dical history was taken, and routin
`I ctrolyte , and
`blood t sts (full blood count, ur a and
`ion te ts}, nasal endoscopy and skin prick t sts
`liver fun
`to common aeroallergens were performed (including
`mixed gra s, positive and negative controls}. Participants
`stopped any usual therapy with decongestants, antihista(cid:173)
`t roids, and wer
`min s, anti- leukotrien s and nasa l
`given sodium cromoglicate nasal spray and eye drops as
`rescue medication . Use of rescue medication was avoided
`for 24 h before each visit.
`Participants attended visit 2 after 2 weeks without their
`usual treatments to establish basel in measurements.
`For the remainder of the study, they took two sprays
`each side (200 meg total} every morning of fluticasone
`nasal spray and either placebo or 5 mg levocetirizine
`tablets. An independent pharmacy encapsulated both
`tablets in an identical manner to blind the study. Flutica(cid:173)
`sone and levocetirizine doses were chosen to represent
`routine clinical practice. A crossover design was used; so
`all ubj ects received 2 weeks of combination therapy with
`fluticasone and levocetirizinc and 2 weeks of monothcr(cid:173)
`apy with fluticasone (and placebo) in a randomized order.
`Visit 3 was conducted after t.he first treatment and visit 4
`after the second.
`
`Measur menls
`
`All outcom s were measured or calculated for baseline
`(visit 2) and aft r ach tr atm nt p riod (vi its 3 and 4}.
`
`Juniper mini Rllinoconjunctivitis Quality-of- Life Ques(cid:173)
`tionnaire. The mini-RQLQ [6] is a validated shortened
`version of th Juniper rllinoconjunctivitis quality-of-life
`questionnaire [7]. There are 14 questions in ftve domains
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`MEDA APTX02173910
`
`PTX0012-00002
`
`CIPLA LTD. EXHIBIT 2024 PAGE 2
`
`

`

`67 8 M. L Barnes ct ol
`
`, ye and other symp(cid:173)
`(activities, practical probl ms, no
`toms]. Each question is scored for the preceding week as
`an integer from 0 (not troubled} to 6 ( xtremely troubled}.
`The global mini-RQLQ score is the average (mean) of all
`question sco res.
`
`Domiciliary morning peak nasal inspiratory flow rate. PNlf
`measurements were noted each morning in a subject diary(cid:173)
`each taken as th b
`t of thr ' from an In -Check'~'.>! PNIF
`meter (Clement Clarke International Ltd, Harlow, UK). This
`is analogous to a reversed peak t1ow meter connected to a
`face mask to establish the maximal airflow rate on fo rced
`nasal inspiration. Subjects were instructed in the correct
`method, and the technique was assessed at scr ening, to
`n ur a eat d postur , horizontal positioning ofth - m ter,
`correct restoration of the reading to zero, a closed mouth,
`and an ad -quat mask seal whil making a maximal nasal
`inspiration.
`
`Domiciliary morning total nasal symptoms score. TNS
`scores w re also recorded in the diary each morning. The
`TNS score is the sum of scores for nasal run, blockage, itch
`and sneeze, each measured on an interval scale of 0, 1, 2
`or 3 rep re enting no symptoms, mild, moderate or severe
`symptoms, re pectiv ·ly. This results in an int g s or for
`TNS of 0 to 12.
`
`Nasal nitric oxide levels. Nit ric oxid
`level ar an objec(cid:173)
`tive marker of airway eosinophilic inflammation [8].
`A Niox ''f' nitric oxide ana lyzer (© 2000 Aerocrine AB,
`Solna, Sweden) was used at each visit to sample nitric
`oxide levels, using a method consistent with the joint
`statement [9] of the American Tho racic Society and the
`European Respiratory Society.
`
`Statistical analysis and data p1·esentation
`The study was powered (a t > 900/o} to detect (P < 0.05) a
`0.7 U change (the MOD} in the primary outcome variable
`- the mini-RQLQ. Th within subject standard d viation
`used was 0.32, as calcul ated in the instrument's init ial
`validation [6]. Analyses were perforn1 d using Minitab,
`Copyright © 2004, Minitab Inc. PA, USA and SPSS for
`Windows (vii} Copyright © 2004, SPSS Inc. Chicago, fl,
`USA.
`Each outcome was assessed for normality usin g Shapir(cid:173)
`o-Wilk tests an.d by eye, with consideration of prev ious
`data s ·ts and the lit rature. All ou comes wer con id red
`normally distributed.
`Fo r PNIF and TNS, analy is was condu t d on mean
`measurements fort.he hnal w ek of each p riod (treatment
`or run- in]; for al l other outcomes single (v isit based}
`measurements were used.
`testing, differences with in
`Before non-superiority
`group (vs. baseline) were tested to demonstrate efficacy
`
`for all outcomes (null hypothesis - in the wider populatio n
`no treatment effect aJsts). Non-superiority testing was
`then conduct d for levoccti riz in vs. placebo add-on
`therapy (null hypothesis -
`in the wider population the
`addit ional beneftt of levocetirizine treatment vs. placebo
`is g reater than the MCID).
`Confidence intervals were calculated for both the e
`two-sided 950/o conftdence intervals for
`comparisons -
`fficacy and on - ided 950/o onftd enc
`intervals for non(cid:173)
`superiority.
`Finally, the number needed to treat (NNT) for one
`subj ect to experience a benefit greater than the MCID for
`each outcome was calculated.
`
`Min imal clinically important diJ]erence determination
`
`The MOD for the min i-RQLQ (± 0. 7 U} was determined in
`its development by an an hor-ba ed approach . MC!Ds fo r
`PNlF (± 6.23), TNS (± 0.52) and nasal nitric oxide (± 68. 7)
`were alculated u ing a di. tribution-based approach -
`each MOD is one-fifth of the outcome's standard devia(cid:173)
`tion at baseline (see Table 3). Anchor- and distribution(cid:173)
`based approache are described in the discussion.
`
`Data presentation
`
`A summa ry by treatment period for each measurement
`taken is given in Tab le 2. Treatment effects for a ll out(cid:173)
`comes (Table 3) were plotted (Fig. 3) using a scale on
`which + 1 U represents an improvement of one MOD, thus
`enab ling different outcomes to be plotted together and
`interpretations to be made of the statistical and clinical
`significance of treatment effects within and between
`group . Tb, plots also al low a compa ri so n of the strength
`of signa l and noise for all outcomes used. Interpretation of
`these plots i described in the di cussion and in Fig. 5.
`
`Results
`
`Patient demographics (see Table I]
`
`Thirty-one subjects were initially enrolled for the study.
`Four subjects chose to withdraw for personal reasons.
`Eleven men and 16 women with mean (SO) age of 45.9
`(15.0) a nd 44.2 (15.9) year
`resp ctively complet d per
`protocol. Sixteen subj ects received fluticasonc with levo(cid:173)
`cetirizin
`(combination) follow d by fluti casone with
`pia ebo (mo notherapy), 11 th opposite. Adverse event
`were reco rd ·d a
`(by tr 'atment period} I x minor epis(cid:173)
`taxis (during combination therapy), 1 x URTI and 1 x
`lethargy (during monotherapy}. No serious adverse events
`occurred.
`
`© 2006 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 36 :676-684
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`MEDA APTX02173911
`
`PTX0012-00003
`
`CIPLA LTD. EXHIBIT 2024 PAGE 3
`
`

`

`levocetirizinc as add-o n -hcrapy to fluticasone in AR 679
`
`Table 2. Means {and standard errors) for all outcomes at baseline and
`following each raodomised tr atmeot
`
`Outcome (u oils)
`
`BaseJi ne
`
`Moooth l"ra py
`
`Combinatio n
`
`1' tV t"'o
`Predicted
`
`Mini-RQLQ (units)
`PNTF (L/min)
`TNS {units)
`Nasal NO {ppb)
`
`2.5 {0.22)
`118 {5.8)
`4.7 {0.46)
`810 {64)
`
`).7 (0.2 5)
`130{5.7)
`2.6 {0.5 1)
`763 (77)
`
`1.6 (0.24)
`l
`I {6.3)
`2. 5 {O. S3)
`758 (67)
`
`91
`
`Mini-ROLQ. nlin.i Rhiuoconjunctivitis Quality-oF-LiFe Questionnaire;
`PNlF, peak nasa l inspiratory fl ow; TN , total mtsal symptoms score; NO,
`nitric oxide.
`
`Tabl 1. Demographics
`
`Particitlanl
`
`Age
`
`Sex
`
`Sensitivibes
`
`Mini-RQLQ
`
`2
`3
`4
`5
`6
`7
`8
`9
`10
`ll
`12
`1J
`l4
`15
`16
`17
`18
`19
`20
`21
`22
`23
`24
`25
`26
`27
`28
`29
`30
`3 1
`
`J l
`63
`55
`37
`48
`55
`33
`28
`66
`47
`60
`32
`57
`58
`65
`23
`61
`60
`0
`29
`44
`49
`18
`2:>
`49
`54
`27
`18
`2 1
`57
`39
`
`F
`M
`F
`M
`M
`M
`F
`M
`M
`F
`F
`M
`F
`F
`F
`F
`F
`M
`M
`M
`F
`M
`F
`F
`F
`F
`F
`F
`F
`M
`F
`
`GHC
`GH
`G
`GTW
`GWHAF
`Gil
`G
`GTWI-IDC
`GWHAF
`GTC
`Gl·l
`GTW
`GWA
`GWH
`GH
`GWHC
`GWH
`GHC
`GTWHCO
`GTWHF
`GDC
`GW
`GW
`GWHC
`GTWHAFDC
`GHC
`GTWI-IAFDC
`GHC
`GTW
`GTW
`GWHC
`
`4.21
`1.64
`!.93
`1.50
`4.86
`1.79
`2.07
`2.79
`2.5 4
`1.07
`3.07
`1.43
`1.86
`2.29
`3. 14
`3.50
`1. 50
`1.57
`2.50
`3.36
`1.71
`3.00
`2.43
`4.29
`5.64
`2.71
`1.4 3
`2.21
`4.71
`4.86
`2.71
`
`67
`
`96
`90
`
`101
`
`Sensitivitirs represen1: G, gr~sses; T, trees; W, weeds; 1-J, housr dust mite;
`A, aspergillus; F, feathers m.i.x: D, dog: C, cat. Mini Rhinoconjunctivitis
`Quality-of-Life Qucstio!lllaire {mini-RQLQ) is the score recorded at
`screening. Percentage predicted FEV 1 was only recorded for asthmatics.
`
`'?
`160
`E
`..... 140
`
`40
`
`20
`
`>-
`~
`"0
`(ij
`;;;)
`0
`"0
`:~
`"0 - 20
`E
`
`Baseline
`
`Monolherapy
`
`Combinalion
`
`Fig. 2. Po lle11 ex posures by treatment period. l3oxplols -boxes show 2"d
`and 3'" qua rtil es, whiskers s how mini ma and maxima.
`
`Pollen Exposure {see Fig. 2)
`
`Dundee pollen proftl es have been published .in thisjournal
`previously [ 10-1 2]. The 2004 pollen counts (52m - 3 , SEM
`
`© 2006 Blac kwell Pu blishing Ltd, C/inicol and Experimental Allergy, 36 : 676-684
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`MEDA APTX02173912
`
`PTX0012-00004
`
`Monolllcrapy
`Combinalion
`Difference
`
`Monolhaapy
`ComJ>;nation
`Ditren:ncc
`
`Monothetapy
`Combination
`Difference
`
`Mooolherapy
`Combination
`Difference
`
`-2
`
`I
`
`•
`
`~· 1-;.--J*
`-..Jt I
`I
`I
`~ :t
`I
`I
`--{t
`!
`
`~t
`0
`
`2
`
`I*
`*
`
`miniRQLQ
`
`t
`t
`
`PNlF
`
`TNS
`
`nNO
`
`4
`
`6
`
`8
`
`Improvement pet number of MCIDs
`
`Fig. 3. Outcome improvements for each randomized treatment and
`eli ITercnccs between trcalmems. Results for monothcrapy {llui.icasone
`and placebo), co mbination (llulicasone ami levocetirizine) and differ(cid:173)
`ences (adrl-on brnef1ts) are ploltt>d on a sca lr to depict changes relative to
`minima l clinically important differences {MODs). For wiUtin treatment
`~!Teets two-sided 95% Is are s hown: ldf'ootes sta tistically and clinically
`significant irnprovern ems. •denotes statistically signifu::anl improve(cid:173)
`ments of uncertain clinical significance. For between treatment tliffer(cid:173)
`ences the one-sided 95% Cis are shown: tdenotes non-superio rity of
`combination therapy.
`
`3.6} were not significantly different to previous years
`(P = 0.67). Individual daily pollen exposure is presented
`by tr atment period in Fig. 2.
`
`Within - and between-treatment effects {see Tables 2-4,
`Fig. J)
`
`Juniper mini rhinoconjunctivitis quality-of-life question(cid:173)
`naire. Statistically
`ignif1cant improvements were een
`fo r change
`from bas lin
`for both monoth rapy
`(P < 0.0001} and ombination therapy (P < 0 .0001}. 1be
`mean (one- ided 95% CI} for th diffi r n
`betw n
`combination and monotberapy (i.e. levo etirizine add-on
`effects) was - 0.1 1 (to - 0.30}, which excludes any benef1t
`greater than the MOD, so we can di miss the null hypoth(cid:173)
`esi and conclude non -superiority: in the wider popula(cid:173)
`tion, when used as add-on to fluticasone nasal pray, the
`
`.. !!! 120
`X ..
`.. 15
`
`:>
`"' 0
`
`Q.
`
`Q.
`
`100
`
`80
`
`60
`
`CIPLA LTD. EXHIBIT 2024 PAGE 4
`
`

`

`680 M. L Barnes ct ol
`
`Tab! 3. Baseline data, MCIO calculations and change by tTearment group for all outcomes
`
`Ou · ome (U)
`
`McnLRQLQ (U)
`PN lF (I../ min)
`TNS( U)*
`
`nNO (pph)
`
`Baseline Mean
`(95%(1)
`
`SD at
`
`b~ scline
`
`MCW
`(S0/5) Monother;Jpy
`
`Combination
`therapy
`
`LEVO ad d.-on eiTe ' I
`
`2.46 (2 .03 to 2.90)
`118 ( 107 to 130)
`4.56 (3 .61 to 5.50)
`810 (682 to 938]
`
`1.17'
`3 1. 2
`2.56
`344
`
`0.71
`6.23
`0.52
`68.7
`
`- 0.70 (- 1.02 to - 0.39), P < 0.0001
`12.0 ( 1.9 to 22.2). P < 0.05
`-2-02 (- 2.91 to - 1.13), P < 0.000 1
`- 37.4 (- 134.5 io 59.6], P= 0.44
`
`- 0.82 (- 1. 15to - 0.49). P < 0.0001
`12.6 (2.6 lo 22.6). P < 0.05
`- 2.13 (- 3.04 to - 1.23), P < 0.0001
`- 43.1 (- 142 to 55). P=0.37
`
`- 0. 11 (to- 0.30). P < 0.0001
`0.57 (to 5.23). P < 0.05
`- 0. 11 (to - 0.51),P < 0.05
`- 5.6 (to - 65. Jl. P < 0.05
`
`Effect da ta are presented as means (95% conlldence intervals] and P-values. P-val ues are the probability of non-efficacy vs. baseline except for add-on
`effects where P-values are the probab i.lity of superiority for combination therapy of 1 minimal clinically irnportam difference or more.
`~ I n dividual symptom score data are presented in Tabl e ~.
`·<calculated iu the instruments deveiOJlment:.
`u:vo, S mg rlai ly levocetirizine Oihyrlrochl orirle; Mini-RQLQ, mini Rhinoconjunctivitis Quality-of-Lif<' Questionnaire; M ID, miJlimal clinically
`importa nt cUfference; PNI F, peak. nasa l inspiratory now ; TNS, total nasa l symptoms scor ; NO, nitric: oxid .
`
`additional min i-RQLQ benefit of levoc:etir.izine tr atment
`vs. placebo is not clinically important. Of27 subjects, two
`experienced a benefit with levoc:etirizine add -on that was
`greater than the MCID - NNT 14 (950/o Cl 5-49).
`
`Domiciliary morning peak nasal inspiratory flow rate. Statis(cid:173)
`tically ign.i&cant improvements were seen for change from
`ba eline for monothernpy (P < 0.05) and combination ther(cid:173)
`apy (P < 0.05). Th m an (one-sid ·:d 950fo Cl) for the
`difterenc: betwe n omb.ination therapy and monotherapy
`(i .. levoc: ti.rizin ' add-on effi cts) wa 0.57 (5.23), whl h
`excludes any ben f1 greater than th MClD, so w can
`dismiss the null hypothesis and conclude non-superiority:
`in the wider population, when used as add-on to tluticasone
`nasal spray, the addjtional PNlF benefit of levocetirizine
`treatment vs. placebo is not clinically important Of 27
`subjects, eight experienced a benef1t with levocetirizine add(cid:173)
`on greater than th · MCID: i.e. NNT 4 (950/o Cl 3 to 7).
`
`Domicilia ry morning Total Nasal Symptoms score. Statis(cid:173)
`tically significant improvement were seen for change
`from baseline for monotherapy (P < 0.0001) and combi(cid:173)
`nation therapy (P < 0.0001). Them an (one-sided 9501o Cl)
`for the difference between combination therapy and
`monotherapy (i.e. 1 vocetirizi ne add-on effects) was
`- O.Jl (- 0.51), which e:xclud es any ben fit greater than
`the MCID, so we can dismiss the null hypothesis and
`conclude non-superiority: in the wider population, when
`used as add -on to tluticasone nasa l spray, the ad ditio.nal
`TNS benefit of levocetirizine treatment vs. placebo is not
`clinically important. Of 27 subjects, n ine experienced a
`benefit with levocetirizine add-on greater than the MClD :
`i .. NNT 3 (95% Cl 2-6). A breakdown by individual
`symptom sco res .i given in Table 4.
`
`Nasal nitl'ic oxide levels. No statistically significa nt im(cid:173)
`provements were seen for change from baseline for
`monothe.rapy
`(P = 0.44)
`or
`co mbination
`therapy
`(P= 0.37). Thus, we were unable to demonstrate a signal
`
`for na al nitric oxide mea urements at all, so no further
`con lu ion should b drawn.
`
`equence analysi
`
`A com parison of randomiza ion g roups showed no stati -
`tically significant differences for carry-over (P= 0.58) or
`period (P = 0.29) effects on the primary outcome.
`
`Discussion
`
`tudy, w s t out to determine wh th ' r
`In th pr s nt
`co mbination therapy with nasal fluticason
`and oral
`levocetirizine was any more effective than fluticasone
`monotherapy. In view of the previous evidence, instead
`of hypothesizing a difference, we set out with the hypoth(cid:173)
`esis that monotherapy could be considered no worse than
`combination therapy, i.e. non-inferior.
`In AR, it is important to mcasur· outcomes not only
`rel ated to nasal airflow obstruction (PNIF, acoustic rhino(cid:173)
`metry or rhinomanometry), but also ymptoms such as
`blockage, rhinorrhoea, itch and sneeze and the effects on
`global quality of life [ 13].
`Both the mini-RQLQ and RQLQ have strong discrimina(cid:173)
`ti ve and valuative mea ur •m nt propertje (defined by
`Guyatt ct at. [14)), and the mini- RQLQ is signifi cantly
`mo re responsiv than the ea rlier rhinoconjunctivitis qual(cid:173)
`ity-of-lite questionnaire [6, 7).
`We have previously shown that PNIF is a rep resentativ
`and repeatable (cv. SOfo) [15] measure of nasal airflow
`obstruction, and that it is more sensitive than acoustic
`rhinometry for monitoring response to histamine nasal
`chall ng s [16]. Na a! nitri oxid
`is consider d a r pro(cid:173)
`ducibl e (cv. lJCVo [17)), non-invasive and easily obtained
`nasal inflammatory marker.
`For our drug interventions we chose to tudy levo etir(cid:173)
`izine ad d-on to fluticasone nasal spray over a 2- week
`period to represent the current best treatment. Fluticasone
`is among the most commonly used of the nasal
`teroids
`available
`in our rh.inology clinic, and has a high
`
`© 2006 Blackwell Publishing Ltd, Clinical and Experimental Allergy, 36:676-684
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`MEDA APTX02173913
`
`PTX0012-00005
`
`CIPLA LTD. EXHIBIT 2024 PAGE 5
`
`

`

`Levoce·irizinc as add-on therapy to fluticasonc in AR 681
`
`Tabl 4. Baseli ne data, MCID calculations and change by treatment group for individual symptom scores
`
`Baseline
`Out·couie Mean
`(U)
`(95%0)
`
`SO at
`baseline
`
`MCJJl
`(SD/ 5) Monotherapy
`
`Combination tberapy
`
`BN( U)
`
`RN IU)
`IT(U)
`
`SN IU)
`
`1.24(0.95 10 1.54)
`1.00 (0.70 0 1.29)
`1.16 (0.86 to 1.46)
`
`1.3 1( 1.04 !0 1.571
`
`0.79
`0.78
`0.79
`
`0.72
`
`0.16
`0.16
`0.16
`
`0.14
`
`- 0.60 ( - 0.63 to - 0.10). I' < 0.0005
`- o.J 7 ( - 0.63 to - 0. 23). P < O.Ot
`- 0.55 ( - 0.85 to - 0.2 5). P< 0.00 1
`- 0.51 ( - 0.79 to - 0.23 ). p < 0.001
`
`- 0.49 ( - 0.74 to - 0.24). P < 0.0005
`- 0.40 ( - 0.66 to - 0. t4). P < o.oos
`- 0.58 ( - 0.88 to - 0.28). P < o.oo 1
`- 0.66 ( - 0.96 to - 0.36). p < 0.0005
`
`LEVO add -<>ll
`o. 11 (to - 0.04). P < 0.005
`- 0.04 (to - 0. t9). P=O. I
`- 0.03 (to - 0. 16). P - 0.06
`
`- O. l5(to - O.J l).P~0. 54
`
`Effect data a re presented as means (9'>% con ft.dence intervals) and P-values. P-values are the probabi[;ty of non-efficacy vs. base[;ne except fo r add-on
`effecrs where P-values are the probability of uperiority for combination therapy of I minimal cl in ically inrponam difference or rnore. EVO = mg
`dai ly levocefjrizine rlihydrochl oride; BN, b lo ck ~ge; RN, rhinorrhoea ; IT, Itching; SN, sneezing; MOD. min imal clinically important difference.
`
`glucocorticoid potency and binding affinity [18, 19],
`wh ile I vocetjriz iw was cho n as pot n ially the most
`effective [20- 23] and Ionge t acting 'new generation' [24]
`antihistamin with a n gl igibl
`potentia l for cardiac
`toxicity at standard doses [13].
`It is not possible in any study to demon trate that a
`treatment has absolutely no effect or that there is abso(cid:173)
`he
`b twe n treatments, becaus
`lutely no differen
`breadth of the confidence interval for such differences is
`only 0 when the sample size is infrnite: for this reason, in
`contrast to efficacy studks, studies hypothesizing equiva-
`1 n
`or non-superiority must f1rst d fmc limits within
`which any change observ d can be considered irrel vant.
`For exam pi , a
`pting that a chang' in PN'IF of I ss than
`± 2 L/min is of no clinical rei vane : 2 L/min is then the
`MCTD for PNTF. An MCTD can be defined as ' tile smallest
`difference in score in the domain of interest which patients
`would perceive as beneficial and which would mandate, in
`the absence of troublesome side-eJfects and excessive cost,
`a change in the patient's management' [25].
`Exactly how a value hould be dct·rmined for an MCID
`is a matter of debate. Perhaps most commonly, it is
`decided on the merits of 'clinical experienc ··, with a
`distinguished writer simply d.ictating what he/sh feels is
`
`-7 - A very gt-eat deal worse
`-6 - A great deal worse
`-5 - A good deal wot-se
`-4 - Moderately wot-se
`-3 - Somewhat worse
`-2 - A little wot-se
`-1 - Almost the same, hardly any wot-se at aU
`o change
`0 -
`+ 1 - Almost the same, hardly any bettet• at all
`+2 - A little bettet·
`omewhat better
`+3 -
`+4 - Moderately better
`+5 - A good deal better
`+6 - A great deal better
`+ 7 - A very gt-eat deal better
`
`Fig. 4. Global r~ting of change questionnaire, rcprod.uc~d by kind
`pennission of Prof. E. Juniper. 'Since your last visit, h~s t·here been any
`change i11 ~clivity limitation/symJll.orns/emotions/overall qualiiy of life,
`related to your aUcrgic rhinitis?' Use the scale shown above.
`
`© 2006 Blackwell Publishing Ltd, C/inicol and Experimental AJ/ergy, 36 : 676-684
`
`of relevance or otherwise. Two more objective method
`hav b en d velop d. In an 'anchor-bas d' approach,
`changes in an outcome of interest arc a e ed and
`compar d with 'a global or overall qu s ion m asuring
`wellbeing or treatment etfect' [26]. For example, in devel(cid:173)
`oping their mini- RQlQ [6], Juniper tal. used a global(cid:173)
`rating scale as illustrated in Fig. 5. They found that when
`assessi ng th global-rating cal alongsid change in the
`min.i-RQLQ, the mean change in their quality-of-life
`instrument for tho e patients that described
`ither a 1 or
`2 U change (positive or negative) on the global scale was
`0. 7 U, and thus stabli h d th MClD of th m ini-RQLQ.
`Th second, more ommonly used 'distribution-ba ed
`approa h' compare magnitud of th chang s en with
`the variability of the measure in question. For example,
`if th '
`·tandard deviation of PNTF measurements in the
`popu lation in general is lO l/min, we can take a propor(cid:173)
`tion of this to represent our MClD. Cohen [27] defined a
`'small change' (which has been interpreted as an MCID) as
`the sample standard deviation of the measure at
`0.2 x
`baseline.
`Once an MClD has been used to set the limits of
`clinically irrelevant chang ·s, the confidence interval for
`any change can be compared not only to 0 (as in fftcacy
`testing) but also to + 1 MCID or -
`I MCfD. Figure 5
`illustrates the interpretation of confidence intervals for a
`change in relation to zero and M IDs.
`As a general rule, anchor-based MCIDs are to b
`favour d to the relatjvely empirica l distribution-based
`MC1Ds. for the mini-RQLQ used in our study, an anchor(cid:173)
`based MCTD had been determin ed in its original va lidation
`[6] . Par the other outcomes, we applied the more com(cid:173)
`monly used 'distribution-based' approach. Where anchor(cid:173)
`based MCID are not available, these are certainly to be
`pr ti rred to b st gue
`approa h
`or the r gr ttably
`common conclusions of no difference when a comparison
`with 0 how P > 0.05 - the corr ct conclusion in uch
`scenarios is usually that the tudy was under-powered.
`Using these method , we have gon beyond the ana(cid:173)
`lyses conducted previously and have been able not only to
`test for a difference within or between groups, but also to
`demonstrate its clinical relevance or irrelevance. We frrst
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`MEDA APTX02173914
`
`PTX0012-00006
`
`CIPLA LTD. EXHIBIT 2024 PAGE 6
`
`

`

`682 M. L Barnes ct ol
`
`Clinical relevance illustration
`
`[ I]
`
`[2]
`
`[3]
`
`[4]
`
`I •
`I (a)
`
`I • I (c)•
`I (b)
`
`(e)_
`
`(t)
`
`I
`
`(d)
`
`I • I
`
`-2
`
`-l
`
`0
`
`2
`
`frnprovcrnent in MC!D
`
`Fig. 5. lnterpr tatiou of resu lts with rega rd to minimal clinically im(cid:173)
`portant differences (MC!Ds) and clinical relevance. The illustration shows
`confHlen e interva ls for 6 hypothetical changes plotted on a scale of
`M lOs of change. For exam ple, if the MCID for peak. nasal inspiratory
`!low rate were 2 L/mio, a change or 4 L/rn in would be plotied at 2 on the
`sca le. Cha nges to the right of zero arr improvemen ts; changes to the left
`are deteriorations. Regions [lj and [4] contain cl inically signiftcanl
`deteriorations and improvements respectively, while Reg ions 12] and [J]
`contain clinically insigniftcant deteriorations and improvements respec(cid:173)
`tively. Thus, confidence i.!llervals (a) and (d) demonstrate effects (d te(cid:173)
`rioration and improvement, respectively) that are both statistically and
`clinically significam: they not only exclude zero e!Tect, but also any
`effect smaller tha n the MCfD. In contrast, co nfidence iJi terva ls (bla nd (c)
`demonstrate effects (deterioration and improvemem, res pectively) which
`although statistica lly signiftcant, are of no cl inica l relevance: th ey are
`I MCID. Conlidenre interva ls (e) aod (f)
`enti rely cootainecl with in
`illusrrare non-sup eriority and non-inferiority, respectively.
`
`demonstrated not only that randomized treatmem with
`tlutitasone alone or in combination with levocetirizine
`producd a stati tkally significa nt improvement from
`baseline (950/o confidence interval excluded 0) but also
`that this improvement was clinically relevant (mean was
`greater than I MCID}. Further, there were no statistically
`ignificant differences between randomized treatments,
`and non-superiority of levocetirizine vs. placebo was
`d monstrated wit}! r ·sp ct to MCIDs for mini -RQLQ, PN IF
`and TNS (one-sided 95% confidence intervals excl ud ed
`any beneftt of J MCTD or more).
`Mean improveme