Record of Proceedings: Therapeutic Products Directorate Scientific Advisory Committee on Respiratory and Allergy Therapies (SAC-RAT) – March 14, 2012

Record of Proceedings

March 14, 2012

Committee Members Present:

Irvin Mayers (Chair), Don Cockcroft, Myrna Dolovich, Allan Donner, Alan Kaplan, Redwan Moqbel, Parameswaran Nair, William Racz, Robert Schellenberg, Matthew Stanbrook, William Swan, Susan Waserman

Via teleconference:

Don Sin

Regrets:

Sharon Dell

Health Canada Representatives:

Office of Science: Erica Harnett, Larissa Lefebvre, Bob Li, Yongmin Yu
Bureau of Cardiology, Allergy and Neurological Sciences: Kimby Barton, Jatinderpal Raj Deol, Gail Grant, Dominique Heon, Ajaykumar Thaker, Violina Thomas, Jun Zhang

Acronyms used in this record:

BCANS:

Bureau of Cardiology, Allergy and Neurological Sciences

BPSIP:

Bureau of Policy, Science and International Programs

CI:

Confidence interval

DPI:

Dry powder inhaler

eNO:

Expired nitric oxide

FeNO:

Fractional exhaled nitric oxide

FEV1:

Forced expiratory volume in one second

ICS:

Inhaled Corticosteroids

MDI:

Metered Dose Inhaler

PC ₂₀:

Provocative concentration causing a 20% fall in FEV1

PD:

Pharmacodynamics

PEF:

Peak expiratory flow

PK:

Pharmacokinetics

R:

Reference product

SAC-RAT:

Scientific Advisory Committee on Respiratory and Allergy Therapies

SE:

Sputum Eosinophils

SME:

Subsequent market entry

T:

Test product

TPD:

Therapeutic Products Directorate

Opening Remarks (K. Barton, BCANS)
Kimby Barton, Director of Bureau of Cardiology, Allergy and Neurological Sciences (BCANS), opened the meeting on behalf of the Director General of the Therapeutic Products Directorate (TPD) at Health Canada. Ms. Barton welcomed the members and thanked them for their time and effort in preparing for and participating in this meeting. Ms. Barton introduced the new SAC-RAT Patient Representative, Mr. William Swan.

Introductions, Affiliations and Interests Declaration (I. Mayers, Chair)
The Chair led the committee in round-table introductions and a review of the agenda. Members introduced themselves, provided brief summaries of their backgrounds and indicated any potential conflicts of interest. No conflicts were identified and all members were allowed to participate fully in the discussions.

At this time, Dr. Moqbel asked that the committee consider replacing him as a core member of the SAC-RAT. Dr. Moqbel suggested being replaced on the committee by a colleague from the University of Alberta.

The Chair acknowledged the passing of Dr. Freddy Hargreave, a member of the SAC-RAT. Dr. Hargreave had an important leadership role in Canadian Respiratory research. His death is a major loss for this community and for the SAC-RAT.

Terms of Reference

The Chair noted that there had been no changes to the Terms of Reference since the last meeting.

Introduction of issues (A. Thaker, BCANS)
Dr. Thaker from the Allergy and Respiratory Drugs Division at BCANS welcomed members and thanked them for their interest in the development of Health Canada's Draft Guidance Document: Data Requirements for Safety and Effectiveness of Subsequent Market Entry Inhaled Corticosteroids Products for Use in the Treatment of Asthma. Dr. Thaker described the two question documents that the committee members were provided prior to the meeting. The first document (document 1) included a set of nine questions, which summarizes the issues/concerns raised by stakeholders during the web consultation process from the fall of 2011. The second document (document 2) included all comments received from stakeholders during this web consultation process.

Presentation from the Bureau of Policy, Science and International Programs (B. Li, BPSIP)

Health Canada made a short presentation that examined a publication by Kelly et al (2006)^{Footnote 1}.

During the web consultation period, stakeholders had expressed concerns regarding the use of sputum eosinophil (SE) counts as the primary end-point for showing therapeutic equivalence of a subsequent market entry (SME) Inhaled Corticosteroid (ICS). A Stakeholder referred to the above-mentioned study (Kelly et al (2006)^{Footnote 1}), which investigated whether a dose-response to inhaled corticosteroids could be shown using SE counts. The Sponsor was concerned that in this study, which measured SE counts in a way similar to that proposed in the draft guidance document, SE counts could not detect a dose-response to ICS. Health Canada critically reviewed the statistical methodology and results of this publication and confirmed the Sponsor's concerns. However, similar results were found for other parameters, such as Forced Expiratory Volume in one second (FEV1), fractional exhaled nitric oxide (FeNO), and asthma symptoms also examined in the study. Only the methacholine provocative concentration causing a 20% fall in FEV1 (PC₂₀) showed a dose-response. Additional publications on the same subject^{Footnote 2,Footnote 3}, proposed by the Committee Chair prior to the meeting were also included in Health Canada's review. The Committee members discussed the presentation and the issues that it highlighted. These discussions are reflected in the specific comments to the questions in Document 1 listed below.

The presentation is available upon request.

Discussions during and after the presentation touched on the following topics:

• The scale for PC₂₀ from the Kelly et al (2006)^{Footnote 1} publication;
• The use of sputum eosinophil counts as an outcome measure;
• The requirement for a dose-response study in the guidance document;
• Re-test reliability and Wash out and Carryover;
• Crossover design;
• Conflicts with innovator and generic products;
• Proposed endpoints.

Summary of these discussions:

• Health Canada cannot base their decisions or guidelines solely on the Kelly et al (2006)^{Footnote 1} publication. Although, it does show a dose-response, this publication is not sufficient to confirm that SE is a reasonable outcome measure.
• Health Canada's guidelines need to be specific enough to ensure that the chosen outcome measure/biomarker can be used by pharmaceutical companies in a reproducible manner.
• There is no biomarker that could be used reliably and reproducibly as an outcome measure.
• Studies need to be designed around the correct patient population. An enriched population will allow for more efficient clinical trials.
• There may be little utility in assessing low-dose drugs in mild-asthmatics as these agents are likely to already be at the top of their dose-response curves.  However, it is an ethical challenge to test low-dose drugs in severe asthmatics.
• To date there is no single biomarker in asthmatic patients that shows a good dose-response curve to inhaled corticosteroids.

The following questions were posed and deliberated by the SAC-RAT members. Each is followed by the committee's final recommendations to Health Canada.

Sputum Eosinophil Counts as the Primary Endpoint

1. Is there scientific and/or clinical consensus that sputum eosinophil counts are an adequately validated measure for use as a primary endpoint for establishing therapeutic equivalence of subsequent market entry (SME) inhaled corticosteroids (ICS) for the treatment of asthma, as currently stated in the Draft Guidance Document?  Please comment.

The committee agreed that, although there is some scientific and clinical evidence to support the use of SE to establish therapeutic equivalence of SME ICS for the treatment of asthma, a statistical confirmation is lacking.  Further, at this time, no known biomarker exists for which there is scientific or clinical consensus of adequate validation for establishing therapeutic equivalence of SME ICS in asthma.  Thus, the measure of SE counts may represent a potential non-invasive biomarker that could be used in clinical trials to assess therapeutic equivalence of SME ICS.  However, its use may be problematic due to issues with the dose-response curves in mild asthmatics.  The committee recommended the inclusion of SE counts (or FeNO) as a method to identify a population with airway inflammation [that is (i.e.) an "enriched population"] which is amenable to ICS treatment.

The committee members further agreed that the limited available evidence must be used to make the best decisions based on what was found to be an imperfect data set to show bioequivalence.  In fact, members suggested that the below statement be added to the guidance document to reflect this.

"Markers recommended in this guidance document are based on the best available evidence for use in this context."

At this time the committee also discussed the clinical efficacy criteria.  These are outlined starting on line 324 of the draft guidance document and describe how SE counts, as a primary efficacy endpoint, should be measured [i.e. "...a difference in the mean sputum eosinophil count (expressed as a percentage of the total count) of at least 50% between the active treatments and the placebo treatments will be considered clinically significant."].

The primary outcome of a study of therapeutic equivalence of SME ICS will involve determining the difference between innovator products and SME products and their corresponding confidence intervals.  To do so, a biomarker is required which will not be at the top of its dose response curve and will require a test of assay sensitivity.  For cases where SE counts are chosen as an outcome measure, the mean percent change from baseline in SE counts for both the innovator and SME drugs will need to be calculated followed by the difference between the two means and their corresponding confidence intervals (CI).  The above-mentioned 50% change was included for tests of assay sensitivity and was meant to validate the chosen endpoint if it was at the top of the dose-response curve.

The following, additional question was also deliberated by the committee at this time.  This was a specific question raised by a study sponsor during the web-consultation period.

Should the measurement of sputum eosinophil counts be limited to study centres which have experience in the sampling and analysis of sputum eosinophils?  How many centres are there in Canada?  Internationally?

Committee members confirmed that studies of SE counts have been conducted in at least 60 centres in Europe and 14 centres in North America.  Information on these sites is available in the public domain.

Sites planning to implement this technique should ensure their personnel are properly trained to measure SE counts, which is a trainable, teachable procedure for which standard operating procedures available.

Alternative Primary Endpoints

2. What are appropriate alternative primary endpoints to establish the therapeutic equivalence of a SME ICS with an innovator ICS product for the treatment of asthma [i.e. FEV1, FeNO, PC₂₀ and peak expiratory flow (PEF)]? Please comment.

Several alternative primary endpoint measurements were discussed by committee members including FEV1, FeNO, PC₂₀ and PEF. There was consensus among the members that PEF, which is an insensitive, effort-dependent method, and represents a poor measure of airflow obstruction, should not be used as an endpoint to show therapeutic equivalence.

The committee acknowledged that there was not strong scientific evidence to support a single outcome measure.

Each parameter, as a potential endpoint, was discussed in detail.  Members agreed on the following recommendations for therapeutic equivalence studies of SME ICS:

• Therapeutic equivalence studies of SME ICS should be conducted in an enriched population to ensure that the population will likely respond to ICS. The population to be studied should show evidence of airway inflammation optimally measured as an increase of induced sputum eosinophils, less optimally by FeNO measures. The effect on SE counts should be measured as currently described in the draft guidance document, line 332. For FeNO, the criteria from the American Thoracic Society Clinical Guideline^{Footnote 4} should be implemented (referencing the normal range for expired NO). If FeNO is selected, patients with sinusitis should be excluded from the study population.

The following options for primary endpoint were suggested:

• Using SE (or alternatively FeNO) as the primary outcome;
• Using FEV1 or PC₂₀ as the primary outcome in an "enriched population". If, in a given study, the population is enriched with eosinophils, it's possible that FEV1 may be the most appropriate outcome measure for therapeutic equivalence.

Committee members also agreed that in each of the above scenarios, the SME ICS will need to be tested in such a way as to prove that it behaves as an ICS should. The draft guidance does not need to include a recommendation for tests of assay sensitivity which demonstrate an improvement from baseline (which is expected for an ICS). Instead this document could describe the requirement that a 50% reduction in SE (or a clinical change in FeNo) be demonstrated. This will ensure a) validation that the SME ICS behaves as an ICS would be expected to behave, and b) that a responsive population was chosen (i.e. an enriched population). This would not involve using SE as an endpoint to measure responsiveness, but rather asks sponsors to prove that tests were done in a context where the drug had an anti-inflammatory effect (i.e. there will be the requirement to show measures of inflammation). If no measures of inflammation are changed, then the study would be rejected on the premise that a proper experiment was not performed.

Additional discussions touched on the following topics:

• The responsiveness of Methacholine PC₂₀;
• Allergen challenge model;
• An "enriched population";
• Sputum eosinophil counts and dose-response curves;
• Proportionality of excipients for multiple doses of SME ICS;
• Complexities of testing ICS products.

The following, additional, questions were also deliberated by the committee at this time. These were specific questions raised by a study sponsor during the web-consultation period.

Is there sufficient evidence that the existing ICS products (innovator products) with established efficacy in the treatment of asthma, will consistently result in a 50% reduction in sputum eosinophils?

The committee agreed that there is sufficient evidence that existing innovator products will consistently result in a reduction of 50% or greater of SE counts.

Since the existing ICS products with established efficacy in the treatment of asthma have not been studied extensively for sputum eosinophils, what will be the implications for the submission if the reference product cannot achieve therapeutic efficacy of 50% over placebo?

The committee agreed that the reference product should achieve therapeutic efficacy of 50% over placebo. If initially, tests do not achieve this result, these should be repeated. Also, it should be ensured that the correct patient population has been selected. If the proper patient population has been chosen on the basis of SE counts, the test should not fail particularly because a bolus dose typically produces about a 90% response.

Therapeutic Equivalence - Study Design

3. Should the guidance document provide separate therapeutic equivalence study design recommendations in terms of study duration, population, sample size, clinical efficacy criteria, therapeutic equivalence criteria, for each possible primary endpoint? If so, please comment.

The committee members agreed that broad guidelines on the design of a clinical study should be outlined in the guidance document.

Recommendations made by the committee will apply equally to metered dose inhalers (MDI) and dry powder inhalers (DPI).  It would not be acceptable to have an innovator MDI and a corresponding generic DPI, based on pharmaceutical equivalence requirements.

Design: Committee members agreed that the parallel group design is the most appropriate.  Cross-over design is not recommended.

Population: The population should be defined as patients with proven mild asthma and evidence of active inflammation based on an inflammatory biomarker, such as FeNO or SE (SE counts of > 3%).  All patients should not have had steroid treatment for at least 6-weeks prior to initiating the clinical trial.

Placebo: The proposed studies must be placebo-controlled in order to show a difference from placebo.

Dose: In general, only the lowest marketed dose needs to be tested and that would be true when the excipients of the innovator and the SME ICS are proportional across all doses. However, if the sponsor has multiple doses and the excipient or propellants concentrations are different from the innovator but still within pharmaceutical equivalence guidelines, then they would be required to test at least two of these doses (preferably the high and low doses) to assess dose proportionality.

Duration: The study duration depends on the primary endpoint selected in the study. Overall, for any outcome measure, the duration of the study must be long enough such that a response can be observed.

If the measurement of SE counts is the primary endpoint, the recommendation on study duration already included in the draft guidance document (i.e. 3-weeks) should be followed. Members indicated that, if SE counts were used as the primary endpoint, it would be important to reiterate the issue of SE count recovery by adding the statement: "It is possible that airway inflammation may not recover to baseline levels."

If PC₂₀ (baseline measurement of 2mg/mL or less) is chosen as an outcome marker in an enriched population, a duration of 4-weeks was deemed to be appropriate.

If pre-bronchodilator FEV1 is chosen as the endpoint in an enriched population, the study duration should be a minimum of 4-weeks.

A formal sample size calculation is required for each primary efficacy endpoint selected in therapeutic equivalence studies.

Therapeutic Equivalence: To show therapeutic equivalence of a SME ICS, a study must use patients who are steroid responsive. An inflammatory biomarker could be used to find the enriched population and to show assay sensitivity. Only steroid responsive patients are to be randomised into the study. Endpoints to assess therapeutic equivalence may include a reduction in inflammation such as with SE counts or NO or changes in FEV1 or PC₂₀.

(The above paragraph was modified in a post-meeting teleconference with Myrna Dolovich, Irvin Mayers and Susan Waserman.)

The therapeutic equivalence criteria when SE counts are used as the primary endpoint have been described in detail in the current draft guidance document and were not discussed during the meeting.

The therapeutic equivalence criteria were also reviewed by the committee for the scenario where FEV1 is used as primary endpoint.  The Draft Guidance document outlines the criteria from line 343. The committee members agreed that this would need to be revised for consistency according to other recommendations elsewhere in this document.

If a sponsor chooses FEV1 or PC₂₀ as a therapeutic equivalence outcome, they will need to first show a mean change of at least 12% for pre-bronchodilator FEV1 or specific percentage mean change for PC₂₀. A difference in the mean FEV1 of at least 12% is considered to be the standard definition of reversibility and would be considered clinically significant. For PC₂₀, assay sensitivity can be shown by one doubling dose.

Health Canada inquired whether secondary endpoints should also be investigated. There was little enthusiasm for other secondary endpoints.  Since the trials are of short duration, it was felt that investigating these endpoints is of little value.

eNO: While eNO was not initially recommended as a possible outcome the committee agreed that it could be used in an enriched population to show that the SME ICS has an anti-inflammatory effect. Thus, it was suggested that a study using eNO as a primary outcome should follow the same design as that for SE.  Expired nitric oxide is very sensitive to steroids and thus, expecting a drop of 50% is not unreasonable such that a change of ± 20% for tests of bioequivalence would also be appropriate.

Criteria for the use of SE count remains unchanged as per the Draft Guideline.

The following additional question was also deliberated by the committee at this time.  This was a specific question raised by a study sponsor during the web-consultation period.

Does the current evidence support the required difference in the mean FEV1 (expressed as percentage of predicted) between the active treatments (post-treatment minus pre-treatment, for both test and reference) and the placebo treatment (post-treatment minus pre-treatment) to be 10% or 12%?

As above, difference in the mean FEV1 of at least 12% is considered to be the standard definition of reversibility and would be considered clinically significant.

4. Depending on the primary endpoint, what should the dose requirements be to establish therapeutic equivalence of SME ICS for the treatment of asthma?  Are at least 2 doses for each dosage strength required?

• Please comment on cases where the SME ICS exhibit non-linear dose dependent pharmacokinetics (PK) profiles or the formulations for different strengths are not dose proportional.

The committee agreed that one dose only for each dosage strength should be tested for therapeutic equivalence in the "enriched population", regardless of the primary outcome. Furthermore, if the product has more than one dosage strength, and the excipients of all proposed dosage strengths are proportional, when the lowest dose of the SME ICS is found to the therapeutically equivalent to the innovator, all doses would be considered acceptable. Otherwise, a new bioequivalence study with two doses (the lowest and the highest to show linearity) would be expected. Alternatively, the dosage strengths not essentially similar to the innovator might be considered a new drug. The guidance should highlight that both the innovator and the SME ICS should not differ by more than 5% (total excipients)).

Safety Studies

5. What should be the required studies to establish the safety of SME ICS for the treatment of asthma? Should the studies include:

• Pharmacokinetics studies when the systemic levels of ICS are measurable?
• Pharmacodynamics (PD) studies when measuring the systemic levels of ICS is not feasible?
• A long-term (1 year) safety study?

Committee members discussed that the intent of the guidance document was not meant to guide safety studies, but rather establish criteria for tests of therapeutic equivalence. It is generally accepted that, for generic products, the safety profile was previously established prior to assessing therapeutic equivalence.

Committee members agreed that a one-year safety study would generally not be required if the therapeutic equivalence between SME and innovator products has been established.

The following additional question was also deliberated by the committee at this time.

Should there be a requirement for a PD study or repeat dose PK study to steady state and then measure or a single dose PK study with a dose higher than the highest dose (double or triple the dose) to measure PK parameters in the blood?

The committee members confirmed that draft guidance document currently recommends that the highest marketed dose for the innovator drug be used for PK studies and that healthy volunteers should be included in the PK study, as per the current draft guidance document. If, at this dose, the ICS still cannot be measured in blood at its highest dose, then a PD (adrenal suppression) study could be used.

The committee confirmed that SME ICS were expected to be measurable in blood and the guidance document should specify that the most sensitive instruments (for example (e.g.) tandem mass spectrometry] be used for these measurements.

The members also discussed the inclusion of a safety study within the bioequivalence study in the asthmatic population and agreed that a PK study using the lowest dose (i.e. that used in the bioequivalence studies, which is able to produce blood measurements), would be acceptable. However, if a higher dose is required to be tested, healthy adult males should be included in the PK study.

Paediatric Studies/Spacers

6. Should separate studies be required in paediatric populations for SME ICS for the treatment of asthma? If so, what type of studies (Therapeutic equivalence, PK/PD, with or without spacers)? Please comment.

The committee commented that this question had been discussed at previous meetings and it was agreed that separate therapeutic equivalence studies should be required in pediatric populations. It was recommended that this requirement should remain unless a proper rationale was provided to explain why a separate study was not necessary. Unfortunately, there was no pediatric respirologist (Dr. Dell) in attendance. The committee members felt that any recommendations for pediatric populations should be discussed with Dr. Dell.

Committee members further agreed that the pediatric population consist of many sub-groups and that the pediatric population needs to be defined in explicit terms. The committee members felt that for this guidance document, a pediatric population would be children between 6 and 12 years (yrs) of age (i.e. 6yrs ≤ 12yrs). Studies in children under 6 yrs of age are difficult to conduct; for these children, inclusion criteria are difficult to establish since asthma is hard to diagnose. The committee suggested that if the reference product is indicated in the pediatric population, studies should be conducted in this group, unless the sponsor could provide a rationale to the contrary. It is possible to use the Asthma Predictive Index (eczema, positive family history, and positive skin test) to phenotype these children rather than using physiologic measurements.

The use of valved spacers in pediatric studies of SME ICS was omitted from the draft guidance. Health Canada explained that this was done in error and that the requirement for the use of spacers will be re-inserted into the guidance document. The committee felt very strongly that spacers were a necessary requirement for these studies. The delivery of an investigational SME ICS in a pediatric study should be implemented in the same manner that the drug would be used outside of the study. In Canada, this is by using a spacer. The spacer to be used during the study should be the same spacer as that recommended for the innovator drug, if specified.

The following additional question was also deliberated by the committee at this time. This was a specific question raised by a study sponsor during the web-consultation period.

Are PK/PD studies in different age ranges (e.g. <4yrs, 4-11yrs) required when the PK of the SME ICS differs between adults and children?

The committee discussed the similarities in PK between a child and an adult. Differences may be observed because children inhale the drug differently and may absorb more of it. Further, attention was drawn to the fact that the above question suggests that there would be no PK/PD difference with the innovator drug. If there were a PK difference between adults and children with the SME ICS (but no difference with the reference product), there is either an issue with either the SME ICS drug or the delivery mechanism.

Safety issues in children are of paramount importance. However, PK/PD safety studies may require submitting children to multiple blood draws (of about 3-4mls each). Another, minimally invasive technique was suggested: a one or two point estimate (population kinetics). The committee discussed whether there were other methods that could be used (e.g. 8am cortisol, 24-hour urine collection, spot urine test). If there were simple ways to obtain the data, the committee felt that separate safety studies should be undertaken in pediatric populations, since it was deemed to be unreasonable to assume that a drug was safe in children based on adult studies. It was left to Health Canada to determine the appropriate methods (most sensitive test and the absolute minimum amount of blood required).

Health Canada agreed to explore data in children 6 - ≤ 11 yrs. and to provide it to the committee. It was agreed that this should not prevent the adult guidance from being developed, but would be included in a future guidance for children.

Statistical Issues

7. Should bootstrap methods be used to construct 90% CIs for the ratio of the test and reference treatment effects on mean changes from baseline compared with placebo?

A a stakeholder expressed concern with log transforming of potentially negative values and proposed the use of the bootstrap technique as an alternative approach to calculating a 90% CI. The committee stated that the bootstrap technique was not a good approach as it is computationally intensive and requires a large sample size.

8. For non-log normal data, should the equivalence criterion be changed to 90% CIs of the difference of mean change from baseline between the test and the reference products falling entirely within 20% of the Reference's mean change from baseline?

The draft guidance document currently states: "To demonstrate the bioequivalence of the test (T) product compared to the reference (R) product, the 90% CI of the T/R ratio of mean change from baseline of the primary efficacy endpoint (e.g. sputum eosinophil count, or FeNO or FEV1) should be within 80-125% based on long transformed data.

The committee proposed that:

1. For normal data, the log should not be used.
2. A revisions to the text in the guidance document might be necessary according to the following statement:

The criterion for equivalence will be satisfied if the 90% confidence interval of the difference in absolute mean counts is entirely within ± 20% of the reference's mean.

This would not be precedent-setting, as ± 20% is a generally accepted range in bioequivalence studies.

9. What method of analysis would be used in the event of a negative change in eosinophil counts since the log transformation of a negative value is undefined?

If the change score is negative, a log transformation cannot be done. This can be corrected by taking the standard approach of log(x+k), where k is a constant that is chosen such that the sum of (x+k) is at least one. Also, the revision above would solve this problem as well.

Prior to adjourning the meeting the committee members had additional discussions regarding the proportionality of excipients and the requirements for quality and quantity of ingredients for pharmaceutical equivalence.

Next Steps, Closing Remarks, and Adjournment of Meeting

Committee members were thanked for their valuable time, productive discussions and recommendations. A brief discussion regarding the next meeting was held. Committee members will meet on June 18, 2012 to discuss cystic fibrosis and pulmonary fibrosis.

Meeting was adjourned.

Please note: Following the meeting, the decision was made to postpone the next meeting of the SAC-RAT to fall 2012.

References: