Clinical trial design, nasal allergen challenge
`models, and considerations of relevance to
`pediatrics, nasal polyposis, and different
`classes of medication
`
`Anders Akerlund , MD, PhD,a Morgan Andersson, MD, PhD,b Jeffrey Leflein, MD,c Torben
`Lildholdt, MD, PhD,d and Niels Mygind, MD9 Lund, Sweden, Ann Arbor, Mich , and Horsens
`and Vejle , Denmark
`
`Guest Editor: Peter H. Howanh, BSc (Hans), DM, FRCP
`Allrihution of Sections :
`Clinical trial design: Anders Akerlund, MD, PhD
`Nasal allergen challenge models: Morgan Andersson, MD, PhD
`Pediatric considerations: 1 effrey Leflein, MD
`Nasal polyposis: Torben Lildholdt, MD, PhD
`Therapeutic considerations: Niels Mygind, MD
`
`Clinical trials in allergic rhinitis present several specific
`difficulties. In seasonal pollen-related disease, there are
`variations between subjects in the extent of pollen sensitization,
`individual variations in exposure to pollen even within a set
`area because of lifestyle differences, and variations between
`different areas in pollen counts and weather patterns. Thus,
`large patient numbers are needed in multicenter trials to
`account for such variations when the standard endpoint is
`symptom reporting. Furthermore, a pollen season may be
`relatively short (eg, lasting 6-8 weeks), and the pollen count is
`inconsistent during thi s period. Crossover study designs are
`thus inappropriate, and trials are usua lly conducted with
`a parallel-group design. This fur ther increases the tri al sample
`size as it reduces statistical power. These large patient numbers
`must be recruited over a very short period. Perennial house
`dust mite-sensitive allergic rhinitis presents other problems.
`Although there is less disease variation, it is appreciated that
`symptoms may be induced by nonallergic as well as allergic
`mechanisms because of the nasal hyperresponsiveness. The
`nonallergic symptoms may not be modified by treatments based
`on allergic disease mechanisms. Thus, symptom
`outcomes-although relevant to the patient-may not
`
`From a AstraZeneca, Lund, Sweden; bDepartments of Otorhinolaryngology
`and Clinical Pharmacology, University Hospital of Lund; 'Allergy and
`Immunology Associates of Ann Arbor; dThe Ear Clinic, Harsens; and ethe
`Department of Internal Medicine, Vejle Hospital, Vejle.
`Disclosure of potential conflict of interest: A. Akerlund-none disclosed. M.
`Andersson- none disclosed. J. Leflein is a member of the Speakers' Bureau
`for AstraZeneca, Merck, and Pfizer Pharmaceuticals. T. Lildholdt- none
`disclosed. N. Mygind is a consultant for the Medical Department of Vejle
`Hospital, Vejle, Denmark.
`Received for publication November 15, 2fX)4; revised December 9, 2004;
`accepted for publication December 13, 2004.
`Reprint request" Anders Akerlund, AstraZeneca, PO Box 34, S-22100 Lund,
`Sweden. E-mail: anders.akerlund@astrazeneca.com.
`0091-6749/$30.00
`© 2005 American Academy of Allergy, Asthma and Immunology
`doi: 10.1 Ol6/j.jaci.2004.12.016
`
`8460
`
`adequately reflect the pharmacologic efficacy of the specific
`interventi on.
`To control variability and focus on allergic di sease mechanisms,
`nasal allergen challenge has been used in drug development.
`Single-dose challenges in the laboratory or in a pollen chamber,
`which allow many volunteers to be studied at the same time,
`have proven useful in the evaluation of drugs that afford acute
`symptom relief. However, such challenges incompletely model
`naturally occurring disease, in which the repeated daily
`exposure to allergen modifies the mucosal inflammatory cell
`profile and in particular promotes the epithelial accumulation
`of effector cells. This alters the response to allergen exposure.
`To model this, repeated low-dose daily allergen exposure has
`been used to generate these mucosal changes artificially, and
`early studies suggest that this may be a more valid model for
`the evaluation of anti-inflammatory therapy. However, little has
`been published with this model.
`Different disease groups are associated with their own specific
`issues in clinical trials. The pediatric population, in which
`allergic r hinitis is common, has different requirements for
`education, quality of life evaluati on, and adverse-event
`monitoring; nasal polyposis, because of the nature of the
`disease, requires addi tional means of assessment, such as nasal
`endoscopy and imaging (eg, computerized tomography
`scanning), as well as attention to additional outcome measures
`( eg, the measurement of sense of smell).
`Within clinical trial design, there are important questions to be
`considered in relationship to the therapeutic intervention.
`Should this be given topically or systemically? What are the
`appropriate timing and frequency of medication? Does the
`disease itself modify the treatment efficacy, and does
`combination therapy afford better clinical outcome than single(cid:173)
`modality therapy? These issues are discussed, and the
`influences of current therapies on objective outcome measures
`in allergic rhinitis are reviewed. (J Allergy Clin Immunol
`2005;115:S460-82.)
`
`Key words: Clinical trial design, seasonal rhinitis, perennial
`rhinitis, inlermillenl rhinitis, persistenl rhinitis, nasal allergen
`challenge, pediatric rhinitis, nasal polyposis, Hrantihistamines,
`intranasal steroids, LTRAs
`
`The basic design components in clinical rhinitis trials
`may at a first glance seem fairly constant, but on closer
`inspection there are considerable variations. These reflect
`the need, under separate circumstances,
`to address
`different objectives, such as the effect of therapeutic
`interventions in preventing the development of seasonal
`allergic rhinitis (SAR) as opposed to reducing symptom
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`PLAINTIFFS' 5
`TRIAL EXHIBIT~ .E
`PTX0053
`]
`
`:0
`
`CIP _DYM_00202264
`
`PTX0053-00001
`
`CIPLA LTD. EXHIBIT 2018 PAGE 1
`
`

`

`J ALLERGY CLIN IMMUNOL
`VOLUME 115, NUMBER 3
`
`Akerlund et al S461
`
`Abbreviations used
`CT: Computerized tomography
`LTRA: Leukotriene receptor antagonist
`NPIF: Nasal peak inspiratory flow
`OME: Otitis media with effusion
`PAR: Perennial allergic rhinitis
`SAR: Seasonal allergic rhinitis
`
`expression in established disease. The fmmer may be
`more suitable for the evaluation of a novel intervention
`that would be anticipated to modify tissue cell recruitment
`than as a therapy that has purely symptomatic benefit. The
`study design thus needs to be tailored to the questions to be
`addressed. For example, with an appropriate study design,
`it has been possible to demonstrate dose-efficacy relation(cid:173)
`ships for intranasal glucocorticoids in allergic rhinitis.
`Objective measurements can support the results acquired
`by subjective symptom scores and add value by increasing
`discriminative ability and confirming effects on inflam(cid:173)
`matory activity. Special considerations are necessary for
`different indications, and the existing classification of
`allergic rhinitis into seasonal and perennial is not always
`sufficiently helpful, whereas the classification of non(cid:173)
`allergic rhinitis is recognized to account for a range of
`different nasal disorders with separate underlying patho(cid:173)
`physiologic mechanisms. An awareness of the practical
`implications of clinical trial design and appropriate
`compromises between theory and clinical reality is
`necessary for a successful clinical trial performance. In
`this section, examples of considerations concerning the
`appropriateness of design in relation to the trial objectives
`and some related caveats are discussed. Specifically, this
`chapter starts with a discussion about general aspects
`of clinical trial design, progresses to challenge models,
`considers application in children, and finishes with a
`discussion about nasal outcomes in nasal polyposis and
`with different classes of medication used in the treatment
`of rhinitis.
`
`GENERAL ASPECTS OF CLINICAL TRIAL
`DESIGN FOR RHINITIS
`Patient populations
`Once a clinical hypothesis and the objective are
`established, an appropriate study population needs to be
`determined. The criteria for the population are key,
`because the appropriateness of the population determines
`the validity of the results. The population should be
`a representative sample ti-om a clinically relevant and
`identifiable population, typically identified by a common
`diagnosis. To achieve a conclusive result ti-om a trial
`effectively, the emollment criteria and the objective of
`the study need to be considered in concert.
`
`Characterization: diagnosis and classification
`The term rhinitis strictly refers to an inflammatory
`condition in the nasal mucosal lining. In daily practice,
`
`however, rhinitis is often colloquially applied to a typical
`symptom constellation, characterized by bouts of sneezing
`and nose running, with or without nasal itch or obstruc(cid:173)
`tion, and as such represents a clinical syndrome rather than
`a specific disease entity. In a wider sense of the term, any
`nasal mucosal modifications caused by, for example,
`pharmacologic effects, endogenous hmmonal effects, or
`exaggerated physiological reactions, such as skier's nose,
`are also included under the te1m rhinitis. It is helpful to be
`aware of the shady overlapping borderline areas between
`nasal symptoms that are physiologically appropriate and
`those that cause discomfon and impair function, which
`constitute disease.
`Primarily, rhinitis is classified as allergic or nonallergic.
`Allergic rhinitis is subdivided into seasonal and perennial
`or inte1mittent and persistent, dependent on the duration of
`symptoms. 1 However, the classification into intermittent
`and persistent is relatively recent, and trials are only now
`being undertaken using this classification. It is probable,
`however, on the basis of standard entry criteria and run-in
`periods, that most patients who enter clinical trials
`investigating the effects of regular medication in either
`seasonal or perennial rhinitis have persistent rhinitis. The
`basis for the change in classification from seasonal and
`perennial to intermittent and persistent relates to the
`appreciation that seasonal allergies in some pans of the
`world may indeed be perennial allergens in others.
`Furthermore, many patients have multiple sensitizations
`(Fig I) and thus, although they are sensitized to seasonal
`allergens, their symptoms last longer than exposure to
`a single specific seasonal allergen. For example, there are
`perennial pollens, such as parietaria. ~ This makes a fmmal
`classification into seasonal or perennial rhinitis irrelevant,
`and recognition of intermittent and persistent allergic
`rhinitis is thus more relevant. 1 Secondary factors, such as
`severity of rhinitis, have been added to make the
`classification clinically useful for therapeutic guidance. 34
`Similar approaches can be made to identify appropriate
`patient populations for clinical trial purposes.
`
`Acquiring the population sample:
`confirmation of the diagnosis
`Seasonal allergic rhinitis is a well-defined class of
`rhinitis with a distinct history easily confirmed by tests
`(skin prick, RAST, or provocation tests). Symptoms can
`be related to exposure with pollen counts, and biochemical
`and cellular indices of allergic inflammation can be
`confirmed by intranasal sampling techniques ( eg, by nasal
`lavage or mucosal brushing). However, such an approach
`is not required to make the diagnosis, which is usually
`clear on the history.
`In perennial allergic rhinitis (PAR), the diagnosis is
`based on a history of symptoms after relevant exposure
`and tests to confirm allergic sensitization, but sampling
`to measure exposure is not feasible in large-scale trials.
`In addition, many patients with perennial rhinitis have
`a mixed basis for their disease, with both allergic and
`nonallergic factors, such as structural airway abnor(cid:173)
`malities, contributing to the symptoms. For example,
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`CIP _DYM_00202265
`
`PTX0053-00002
`
`CIPLA LTD. EXHIBIT 2018 PAGE 2
`
`

`

`S462 Akerlund et al
`
`.;t;:{t
`
`h:
`
`,,(X)
`
`·~
`.'~
`
`:t .;!(>)
`~ z
`"-
`
`~~(
`
`FIG 1. Seasonal changes in pollen count illustrating the potential
`effect of variation exposure on clinical t ria l outcome if studies are
`undertaken for 2 weeks at early in season (A), at peak season (B),
`and late in season as pollen count falls and symptoms spontane(cid:173)
`ously improve (C).
`
`a retrospective analysis of 975 rhinitis patients in the
`United States found that 23% could be classified as having
`pure nonallergic rhinitis, 43% pure allergic rhinitis, and
`34% mixed rhinitis. 5 Accordingly, the magnitude of
`improvement with a therapeutic intervention in allergic
`rhinitis might be greater in those with pure disease than
`those with mixed disease. Thus, the majority of clinical
`trials of therapies in allergic rhinitis have focused on
`seasonal allergic disease.
`Perennial nonallergic rhinitis is a diagnosis in which
`positive diagnostic criteria are often lacking. Subgroups
`are often insufficiently characterized even if discrimina(cid:173)
`tive indicators can identify a distinctive subgroup. For
`example, the presence of eosinophilic inflammation on
`intranasal sampling can lead to a diagnosis of nonallergic
`rhinitis with eosinophilia syndrome in subjects with
`rhinitis without specific IgE to common aeroallergens.
`Sensitivity to specific agents, such as aspirin, can
`occasionally be confirmed by provocation tests. Still,
`regardless of extensive efforts, on numerous occasions,
`perennial nonallergic rhinitis will be diagnosed on the
`basis of presence of symptoms in the absence of positive
`allergy tests or conclusive history.
`
`Compromise between scientific theory
`and reality
`Design considerations may lead to conflict between
`scientific ambitions and realities in clinical practice and,
`therefore, a need for compromise. The real-life situation in
`a patient consultation should be applied if the objective is
`to study drug efficacy. Under these circumstances, the
`patient's need for treatment may be the only relevant
`inclusion criterion. In a clinical situation, the history,
`a set of symptoms, and physical signs are sufficient for
`a therapeutic decision. However, this approach is not
`satisfactory in a scientific context in which measurable and
`verifiable data are wanted, and the diagnosis must be
`unambiguously confirmed. Scientific stringency and study
`management according to good clinical practice demand
`more precise diagnostic criteria.
`
`J ALLERGY CLIN IMMUNOL
`MARCH 2005
`
`Assessment of efficacy in clinical trials
`Multiple factors contribute to variability in efficacy
`assessments. These include environmental factors such as
`allergen exposure, patient factors such as adherence, and
`treatment-related factors such as the intrinsic activity of
`the drug. Assessment of nasal symptoms is the mandatory
`clinical efficacy variable for evaluation of drug effects in
`rhinitis.
`
`Symptom severity assessment
`Assessment of efficacy in rhinitis is primarily based on
`subjective grading of symptom severity, as discussed in
`the section on clinical trial outcomes. Because symptom
`assessments are subjective, they are sensitive to factors
`that affect patients' experience of symptoms, including
`expectations, emotions, personality, personal perception,
`and basis of reference. Score step scales, such as rating
`scales or visual analogue scales, can be applied. Multiple
`symptoms are relevant and can be explored, but the mass(cid:173)
`significance phenomenon must be kept in mind-that is,
`the risk that an abundance of variables can lead to chance
`findings and incorrect conclusions.
`To minimize this risk, the number of efficacy variables
`should be kept at a minimum, preferably without com(cid:173)
`promising different aspects of the patient's experience.
`This can be accomplished by using a composite score
`which may consist of the 3 basic symptoms of rhinitis: (1)
`nasal blockage, (2) hypersecretion, and (3) the irritative
`sensory stimulus, presenting as nasal itch, or as its reflex
`response, sneezing. Typically, the 3 symptoms are given
`equal weight. Sometimes itch and sneeze are included
`separately, constituting a 4 basic composite symptoms
`score. With the latter approach, the iiTitati ve sensory
`symptom is weighed to at least 50% of the composite
`score, because rhinorrhea is also a symptom predomi(cid:173)
`nantly determined by sensory neural stimulation. The
`composites of the subjective assessments are not strictly
`independent variables; for instance, secretions may affect
`scoring for blockage.
`The grading of symptoms can be based on severity or
`duration (hours per day). Assessments can be either
`instantaneous or reflective over a defined period (eg, last
`12 hours). The former is necessary when precise time is
`relevant, as in onset or duration of action studies or with
`once-a-day medication, to verify 24-hour duration of
`therapy. 6 For subjective symptom scores, the information
`will always originate from the patient. Assessments are
`made directly by the patient or by the physician, but the
`latter will add filtering to the information.
`It is also relevant to gather a more general assessment of
`efficacy. The global assessment of treatment efficacy and
`symptom relief by patients is a subjective assessment
`based on the patients' recollection and personal reference
`base. These are clinically relevant because they reflect the
`information clinicians use to evaluate efficacy of treatment
`at follow -up visits. The risk of recall bias and the lack of
`comparison with a baseline measure make this a less
`distinct variable. Nevertheless, it is a useful tool for
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`CIP _DYM_00202266
`
`PTX0053-00003
`
`CIPLA LTD. EXHIBIT 2018 PAGE 3
`
`

`

`J ALLERGY CLIN IMMUNOL
`VOLUME 115, NUMBER 3
`
`Akerlund et al S463
`
`assessment of efficacy,7 as is health-related quality-of-life
`(see the section on clinical outcomes).
`
`Objective measurements
`The reliability of subjective assessments is enhanced if
`they can be confirmed objectively. Among the objective
`methods, nasal peak flow offers daily measurements by
`patients at home (see the section on objective monitoring
`of nasal patency). Nasal peak inspiratory flow (NPIF) is an
`objective measw·e of nasal patency that can be used for
`home monitoring to link to the patient's experience of
`nasal congestion and gives a fair estimate of upper airway
`function with a coefficient of variation of around 10% 8
`Differences in efficacy between doses and drugs within
`a narrow interval can be detected by NPIF.9 Other
`methods applicable in a large-scale situation are acoustic
`rhinometry and rhinomanometry (see chapter on nasal
`patency), b ut the added value of these more demanding
`methods over NPIF in large-scale trials has not been
`proven. The time of day and technique of measurement
`should be standardized and patients familiarized
`to
`procedures.
`Nasal inflammatory indices can confirm the anti(cid:173)
`inflammatory effects of an intervention (see section on
`objective monitoring of nasal inflammation).
`
`Pollen exposure
`An obvious external factor is exposure to allergen.
`Total absence of symptoms in pollen-sensitive patients
`with rhinitis out of season is followed by a priming period
`in which symptoms and reactivity to allergen gradually
`to allergen drives
`increase and continued exposure
`symptoms. Monitoring pollen counts in the areas where
`the study is ongoing gives only a very rough estimate of
`individual exposw·e. Pollen exposure can vary consider(cid:173)
`ably, not only from one season to another but also day by
`day and from one place to another. Exposure is largely
`dependent on the weather conditions during the pollina(cid:173)
`tion period but also on individual factors. Patient-related
`factors like frequency and duration of outdoor activities,
`ability to avoid exposure, and individual reactivity will
`affect symptom severity. For example, 2 individuals
`within the same area, with comparable degrees of seasonal
`aeroallergen sensitization, will have different symptom
`severity if one has an outdoor job and the other works
`indoors. Personal monitoring of exposure, although
`possible, is technically complicated and not feasible in
`full-scale trials.
`If the intrinsic efficacy of the drug under evaluation is
`a primary objective, as in dose-finding trials, it is helpful to
`control pollen exposure by only taking into account days
`with a minimum level of exposure. As an example, in
`a seasonal study where no difference between 2 nasal
`steroids was found in the overall analysis, an efficacy
`difference was detected on days with at least moderate
`exposure. 10 This indicates that sufficient exposure may
`enable detection of more subtle differences. The relevance
`of exposure is further illustrated by subanal yses in a trial
`conducted in 2 geographical regions. In one region, the
`
`ragweed pollen season was trivial, as confirmed by daily
`monitoring, and in the other region met the anticipations
`of the triaL In the latter, a significant dose-response was
`found in the subpopulation with low allergen exposure,
`and no significant differences were found within a 4-fold
`range of doses. 11
`Patients receiving effective treatment have a heteroge(cid:173)
`nous response and could have decreasing, unchanged, or
`even increasing symptom severity, depending on circum(cid:173)
`stances and the starting point (Fig 1). Placebo is thus
`necessary to ensure relevant exposw·e. In the absence of
`placebo control, a lack of exposure can lead to false
`conclusions regarding efficacy.
`
`Crossover versus parallel group design
`A crossover design has the advantage that patients act
`as their own conu·ols, and there is thus a need for fewer
`patients by eliminating interindividual variation, under the
`assumption that disease severity is comparable in all
`periods. This is suitable for studies in PAR with contin(cid:173)
`uous symptoms, but not for continuous treatment trials in
`seasonal disease. Although randomization will balance
`differences between periods and sequence effects can be
`dealt with statistically, confounding because of environ(cid:173)
`mental factors, mainly variations in pollen exposure in
`seasonal disease, can complicate interpretation of results
`and possibly eliminate the advantage of this design. The
`length of a crossover trial also introduces an added risk. A
`run-in period lasting from a few days to as long as a week
`is advisable to establish a solid baseline, followed by
`treatment periods lasting 2 to 3 weeks, separated by
`a washout period whose length is dependent on the drug
`under evaluation. Thus, even the simplest comparison of
`an active agent with its placebo will take 6 weeks and may
`endanger the comparability of treatment periods in a pollen
`season lasting 1 to 2 months.
`Treatment periods may be subject to rising or falling
`pollen counts and varying sensitivity to treatment because
`of different stages of inflammation. For all of these
`reasons, there are theoretical advantages to a parallel(cid:173)
`group design, making it preferable for most trials in SAR,
`with the caveat that larger numbers of patients are required
`compared with a crossover study.
`
`Dose-response
`If variation is minimized, the ability to detect the in(cid:173)
`trinsic efficacy of a drug and differences between doses is
`enhanced. PAR is a more heterogeneous population than
`SAR, and exposure to allergen can vary largely. Attempts
`to demonstrate a dose-response in PAR studies, in the
`sense that one dose is statistically significantly superior to
`another, often fail, 12 even if it is occasionally possible. 13
`Through the application of provocation models, such
`as the repeated nasal allergen challenge model to create
`a nasal allergen challenge artificial season, it has been
`possible not only to demonstrate dose responses 14 but
`also to compare efficacy between compounds. 15•16 Such
`models can be helpful to compare the clinical potency of
`different compounds.
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`CIP _DYM_00202267
`
`PTX0053-00004
`
`CIPLA LTD. EXHIBIT 2018 PAGE 4
`
`

`

`S464 Akerlund et al
`
`J ALLERGY CLIN IMMUNOL
`MARCH 2005
`
`Placebo effect and regression to the mean
`A baseline period before start of treatment will offer the
`best reference when the change induced by treatment is of
`interest. It is not unique that the mean symptom severity
`progressively increases the week before start of treatment.
`There is a risk of a filtered population sample. If inclusion
`in the trial requires minimum symptom severity, patients
`scoring high, having bad days could be favored for
`inclusion. Patients included at the peak of symptoms in
`a condition where symptoms may vary spontaneously
`over time are not likely to get worse. The result will be
`a population that will have preponderance toward im(cid:173)
`provement. This improvement, which could be interpreted
`as efficacy, will also be seen in the placebo group, and this
`accounts for some of the improvement seen in the placebo
`group. Thus, for the evaluation of efficacy, comparisons
`with placebo are essential.
`Rhinitis trials, in which primary efficacy variables are
`subjective, are especially sensitive to a placebo driven by
`expectations and bias. Clearly, there are effects of a nasal
`spray placebo, with as much as a 50% change from
`baseline with such therapy .17 An aqueous nasal spray in
`itself is likely to be an active treatment rather than a pure
`placebo, because the mere moisturizing or cleansing effect
`of a nasal spray will dilute or wash out mediators and
`relieve symptoms. However, even with an oral placebo,
`a >30% reduction in symptom scores has been found in
`some studies. 1R A third factor that can contribute to
`efficacy in placebo groups is patients ' use of other effec(cid:173)
`tive medications, either supplied rescue medication or
`undetected use of over-the-counter remedies.
`
`Significance of findings
`The size of the population sample will dete1mine
`statistical ability to detect efficacy. Large and consistent
`clinical effects require smaller study numbers in compar(cid:173)
`ison with placebo than an intervention with a less con(cid:173)
`sistent effect. With appropriate! y large studies, small
`effects can be discerned statistically, but the clinical
`therapy must be questioned.
`relevance of such
`Regardless of whether the objective is to find out if
`a compound is effective at all or if efficacy can reach
`a certain threshold, the clinically relevant efficacy must be
`predetermined. This is a prerequisite for an adequate
`sample size calculation. The magnitude of efficacy-for
`example, nasal blockage changing 1 score step from
`moderate to mild-is easy to relate to an individual. The
`mean symptom score value of a population, however, is an
`abstract quantity, but can be translated to a clinically
`relevant quantity. The appreciation that a 0.5 mean score
`step improvement can be translated to, for example, an
`improvement of at least 1 score step in 50% of a population
`makes this easier to conceptualize. An alternative ap(cid:173)
`proach is to calculate the number needed to treat to achieve
`a particular outcome. Sadly, few clinical trials provide
`information in this format.
`A special requirement for drug registration is to
`determine a lowest effective dose. A common scenario
`
`is that a drug has a defined dose response. Under the
`assumption of a sigmoidal dose-effect curve (Fig 2), any
`dose above the zero efficacy level can be found effective
`(statistically significant efficacy regardless of magnitude)
`given a large enough sample size. The discriminative
`capacity of the efficacy variable and the patient numbers
`will determine what magnitude of ditl'erence is statistically
`significant. The sample size must be driven by the
`magnitude of change seen as clinically relevant for the
`primary efficacy variable. Thus, there is relativity to
`the meaning of effeciive.
`Fortunately, in rhinitis, there is guidance from estab(cid:173)
`lished efficacious reference drugs. A ditJerence between
`u·eatments in conventional clinical rhinitis u·ials of a 0.5
`score step on a 0 to 9 scale (composite of blockage,
`secretion, and sneezing) can be detected with reasonable
`numbers of patients and has been considered a clinically
`relevant magnitude of efficacy. 12·1. 3 The clinical relevance
`will of co w·se depend on the starting reference symptom
`score, because a change from 1.0 to 0.5 might be con(cid:173)
`sidered a more relevant reduction than that from 8.0
`to 7.5, even if both were statistically significant. In(cid:173)
`sufficient attention has been given to the magnitude of
`change in relationship to the baseline scores in the
`presentation of results. This aside, nasal corticosteroids
`can achieve superior efficacy by more than 2 score steps
`compared with placebo. 1 0
`A minimum of 2 doses with statistically significant
`difference in efficacy is necessary to make any clinical
`potency comparison between 2 drugs. This means the
`doses must be on the slope of the dose-effect curve (Fig 2,
`example A) to be able to demonstrate a dose dependency.
`If the doses compared are above the flat top end of the
`curve (Fig 2, example B), any doses could be chosen and
`would still be found not statistically ditl'erent. Of course,
`" no statistical difference " between 2 medications cannot
`be interpreted as equal efficacy. To show equivalence,
`predetermined criteria for equality and a different statis(cid:173)
`tical algorithm are necessary, usually demanding far
`bigger sample sizes.
`
`Practical considerations
`Time constraints typically apply to trials in rhinitis. If
`PAR is being studied, seasonal allergens or postseasonal
`hyperreactivity could influence the symptoms. It is
`necessary to time trials with a sufficient margin to a void
`the major pollen season. In contrast, it is crucial to hit the
`season for SAR (Fig 1 ). The ambition is to include patients
`with comparable state of priming, duration of symptoms,
`and allergic inflammation. Start of treatment for all
`patients within a week or less, and within area region, is
`a way to make the disease under study more consistent. In
`multicenter studies, this may or may not be applicable
`depending on the geographic dispersion of the centers and
`the seasonal pattern of pollen exposure at those sites.
`There is a risk of unfavorable weather conditions with
`low airborne concentrations of pollen. This will lead to
`lack of symptoms or very mild symptoms and reduce the
`
`HIGHLY CONFIDENTIAL-
`SUBJECT TO STIPULATED PROTECTIVE ORDER
`
`CIP _DYM_00202268
`
`PTX0053-00005
`
`CIPLA LTD. EXHIBIT 2018 PAGE 5
`
`

`

`J ALLERGY CLIN IMMUNOL
`VOLUME 115, NUMBER 3
`
`Akerlund et al S465
`
`potential to demonstrate any etlicacy. Including regions
`with reliable and predictable pollen seasons is thus pre(cid:173)
`ferable.
`Preventive treatment is in principle different from
`a situation in which the stmt of treatment is driven by
`presence of symptoms. This distinction can be very
`delicate. Pollen forecasts and day-by-day monitoring of
`counts can help predict factors beyond human control to
`allow fine -tuning of time for start of treatment. To ensure
`that a symptomatic population is enrolled, minimum
`symptom criteria can be applied.
`There are high demands on study management and
`logistics. Recruitment and enrollment of large numbers of
`patients within few days, at short notice, in a lm·ge number
`of geographically spread centers may be necessary.
`Scientific stringency must be weighed against practical
`feasibility. How far this compromise should be taken must
`be driven by the objectives. A ve1y regulated and su·ict
`protocol with an effect on a patient' s daily life can actually
`have an effect opposite the intended stringency. Re(cid:173)
`cruitment will be ditlicult and will lead to a biased
`population s3lllple, and there is a risk that lack of
`compliance will be concealed because of unwillingness
`to report divergence. As with most endeavors, experience
`and planning are key to success.
`
`DISEASE MODELS OF ALLERGIC RHINITIS
`
`One further step toward the study of efficacy of drugs in
`rhinitis is to standardize by controlling multiple factors,
`similar to a classical laboratory experiment.
`Artificial seasonal models can achieve this, as in the
`nasal allergen challenge artificial season developed at
`Lund University, which is a daily repeat challenge model
`that produces