Bioavailability and Bioequivalence Studies for Nasal Aerosols and Nasal Sprays for Local Action

Administrative draft guidance, not enacted legislation. This captured publisher edition preserves the listed document and attachments; later developments are not independently consolidated.

Guidance for Industry
Bioavailability and
Bioequivalence
Studies for Nasal Aerosols
and
Nasal Sprays 	for 	Local 	Action
DRAFT GUIDANCE
This guidance document is being distributed for comment purposes only.
Comments and suggestions regarding this draft document should be submitted within 60 days of
publication of the Federal Register notice announcing the availability of the draft guidance. Submit
comments to Dockets Management Branch (HFA-305), Food and Drug Administration,
5630 Fishers Lane, rm. 1601, Rockville, MD 20857. All comments should be identified with the
docket number listed in the notice of availability that published in the Federal Register.
For questions on the content of the draft document contact Wallace Adams, 301-594-5618.
U.S. Department of Health and Human Services
Food and Drug Administration
Center for Drug Evaluation and Research (CDER)
Biopharmaceutics
April 2003

Guidance for Industry
Bioavailability and
Bioequivalence
Studies for Nasal Aerosols
and
Nasal Sprays 	for 	Local
Action
Additional copies are available from:
Division of Drug Information (HFD-240)
Center for Drug Evaluation and Research (CDER)
5600 Fishers Lane,
Rockville, MD 20857 (Tel) 301-827-4573
Internet at http://www.fda.gov/cder/guidance/index.htm
U.S. Department of Health and Human Services
Food and Drug Administration
Center for Drug Evaluation and Research (CDER)
Biopharmaceutics
April
2003

Contains Nonbinding Recommendations
Draft — Not for Implementation
TABLE OF CONTENTS
I. 	INTRODUCTION............................................................................................................. 1
II. 	BACKGROUND ............................................................................................................... 2
A. 	BA and BE Data ........................................................................................................................... 3
1. Local Delivery BA/BE Concepts. ..................................................................................................... 3
2. Systemic Exposure and Systemic Absorption BA/BE Concepts. ...................................................... 4
B. 	CMC and In Vitro BA Tests (Noncomparative) Versus BE Tests (Comparative) .................. 5
III. 	FORMULATION AND CONTAINER AND CLOSURE SYSTEM......................... 5
A. 	Formulation.................................................................................................................................... 5
B. 	Container and Closure System.................................................................................................. 6
IV. 	DOCUMENTATION OF BA AND BE ........................................................................ 6
A. 	NDAs ............................................................................................................................................... 6
B. 	ANDAs ............................................................................................................................................ 6
1. Solution Formulations. .................................................................................................................... 7
2. Suspension Formulations with PK Systemic Exposure Data. .......................................................... 7
3. Suspension Formulations without PK Systemic Exposure Data. ..................................................... 7
C. 	Postapproval Change ................................................................................................................. 8
V. 	IN VITRO STUDIES ...................................................................................................... 8
A. 	Batches and Drug Product Sample Collection........................................................................ 8
1. NDAs................................................................................................................................................ 8
2. ANDAs. ............................................................................................................................................ 9
B. 	Tests and Metrics ......................................................................................................................... 9
1. Single Actuation Content (SAC) Through Container Life. ............................................................ 11
2. Droplet Size Distribution by Laser Diffraction.............................................................................. 12
a. Nasal sprays. .............................................................................................................................. 12
b. Nasal aerosols. ........................................................................................................................... 13
3. Drug in Small Particles/Droplets, or Particle/Droplet Size Distribution by Cascade Impactor... 14
a. Nasal sprays: Drug in Small Particles/Droplets. ........................................................................ 14
b. Nasal aerosols: Particle/Droplet Size Distribution..................................................................... 15
4. Drug Particle Size Distribution by Microscopy. ............................................................................ 15
5. Spray Pattern. ................................................................................................................................ 16
a. For nonimpaction systems.......................................................................................................... 17
b. For impaction systems. .............................................................................................................. 17
c. For both nonimpaction and impaction systems. ........................................................................ .18
6. Plume Geometry. ........................................................................................................................... 18
7. Priming and Repriming.................................................................................................................. 20
VI. 	CLINICAL STUDIES FOR LOCAL DELIVERY ................................................. 20
A. 	General Information ................................................................................................................. 20
1. NDAs.............................................................................................................................................. 20
2. ANDAs. .......................................................................................................................................... 21
B. 	Clinical Study Batches............................................................................................................... 21

Contains Nonbinding Recommendations
Draft — Not for Implementation
C. 	Clinical BE Study Design and Subject Inclusion Criteria.................................................. 22
D. 	Clinical BE Study Endpoints ................................................................................................... 23
VII. 	PK STUDIES FOR SYSTEMIC EXPOSURE ......................................................... 24
A. 	General Information ................................................................................................................. 24
B. 	Study Batches .............................................................................................................................. 25
C. 	Study Design and Subject Inclusion Criteria ....................................................................... 25
D. 	Study Measures ............................................................................................................................ 26
VIII. 	PD OR CLINICAL STUDIES FOR SYSTEMIC ABSORPTION....................... 26
A. 	General Information .................................................................................................................. 26
B. 	Clinical Study Batches............................................................................................................... 27
C. 	Clinical BE Study Designs and Subject Inclusion Criteria ................................................ 27
D. 	Clinical BE Study Endpoints for Corticosteriods ............................................................... 28
IX. 	NUMBER OF RESERVE SAMPLES FOR BA AND BE TESTING.................... 29
X. 	MULTIPLE STRENGTHS......................................................................................... 29
A. 	Solution Formulation Nasal Sprays........................................................................................ 30
B. 	Suspension Formulation Nasal Sprays................................................................................... 30
XI. 	SMALLER CONTAINER SIZE S ............................................................................... 31
REFERENCES.......................................................................................................................... 31
TABLE 1 ...................................................................................................................................... 32
Note: The following stand alone documents will be provided when completed.
APPENDIX A: DECISION TREE FOR PRODUCT QUALITY STUDIES
APPENDIX B: STATISTICS FOR IN VITRO BA DATA
APPENDIX C: NONPROFILE IN VITRO BE DATA — USING PBE STATISTICS
APPENDIX D: NONPROFILE IN VITRO BE DATA — USING PBE STATISTICS
APPENDIX E: STATISTICS FOR IN VITRO PROFILE COMPARISONS
APPENDIX F: STATISTICS FOR ALLERGIC RHINITIS STUDIES
APPENDIX G: STATISTICS FOR SYSTEMIC EXPOSURE AND ABSORPTION

Contains Nonbinding Recommendations
Draft — Not for Implementation
1
1 	Guidance for Industry
2
3 	Bioavailability and Bioequivalence Studies for Nasal Aerosols
4 	and Nasal Sprays for Local Action
5
6
7 	This draft guidance, when finalized, will represent the Food and Drug Administration’s (FDA’s) current
8 	thinking on this topic. It does not create or confer any rights for or on any person and does not operate to
9 	bind FDA or the public. You can use an alternative approach if the approach satisfies the requirements of
10 	the applicable statutes and regulations. If you want to discuss an alternative approach, contact the FDA
11 	staff responsible for implementing this guidance. If you cannot identify the appropriate FDA staff, call
12 	the appropriate number listed on the title page of this guidance.
13
14
15
16 	I. 	INTRODUCTION
17
18 	This guidance is intended to provide recommendations to applicants who are planning product
19 	quality studies to measure bioavailability (BA) and/or establish bioequivalence (BE) in support
20 	of new drug applications (NDAs) or abbreviated new drug applications (ANDAs) for locally
21 	acting drugs in nasal aerosols (metered-dose inhalers (MDIs)) and nasal sprays (metered-dose
22 	spray pumps). This guidance addresses BA and BE studies of prescription corticosteroids,
23 	antihistamines, anticholinergic drug products, and the over-the-counter (OTC) mast-cell
24 	stabilizer cromolyn sodium. Applicability of the guidance to other classes of intranasal drugs
25 	that may be developed in the future should be discussed with the appropriate CDER review
26 	division.
27
28 	This guidance does not cover studies of nasal sprays included in an applicable OTC monograph2
29 	or studies of (1) metered-dose products intended to deliver drug systemically via the nasal route
30 	or (2) drugs in nasal nonmetered dose atomizer (squeeze) bottles that require premarket
31 	approval.
32
33 	The first draft of this guidance was issued in June 1999 for comment. Because of changes made
34 	as a result of comments received to the docket, internal discussions, and deliberations of the
35 	Advisory Committee for Pharmaceutical Science, we have decided to issue the guidance once
36 	again in draft. A series of attachments are being developed and will be posted with this draft
1 This guidance has been prepared by the Oral Inhalation and Nasal Drug Products Technical Committee, Locally
Acting Drug Products Steering Committee, Biopharmaceutics Coordinating Committee, with contributions from the
Inhalation Drug Products Working Group, the Chemistry, Manufacturing, and Controls Coordinating Committee, in
the Center for Drug Evaluation and Research (CDER) at the Food and Drug Administration.
2 21 CFR 341. Cold, Cough, Allergy, Bronchodilator, and Antiasthmatic Drug Products for Over-the-Counter
Human Use.
1

Contains Nonbinding Recommendations
Draft — Not for Implementation
37 	guidance as stand alone documents on the Internet as soon as they have been completed.
38
39 	FDA’s guidance documents, including this guidance, do not establish legally enforceable
40 	responsibilities. Instead, guidances describe the Agency’s current thinking on a topic and should
41 	be viewed only as recommendations, unless specific regulatory or statutory requirements are
42 	cited. The use of the word should in Agency guidances means that something is suggested or
43 	recommended, but not required.
44
45
46 	II. 	BACKGROUND
47
48 	Product quality studies provide information that pertains to the identity, strength, quality, purity,
49 	and potency of a drug product. These studies include information on chemistry, manufacturing,
50 	and controls (CMC), microbiology, BE and certain aspects of BA. 	A BE study is normally used
51 	to compare a test product (T) to a reference product (R): the to-be-marketed product is compared
52 	to a pivotal clinical trial material, and a generic product is compared to a reference listed drug. A
53 	BE study thus provides information on product quality. BA studies for ensuring product quality
54 	relate to the release of the active ingredient or active moiety from the drug product (Williams et
55 	al., 2000). BA studies may also address biopharmaceutical and clinical pharmacology issues,
56 	such as absorption, distribution, and elimination of the active ingredient and its metabolites and
57 	dose proportionality. These latter BA/PK studies provide information beyond product quality BA
58 	characterization and would also be included in the Human Pharmacokinetics section (Item 6) of
59 	an NDA. These latter studies are not the subject of this guidance. Rather, this guidance discusses
60 	studies that focus on product performance (i.e., release of a drug substance from a drug product).
61 	Subsequent references to BA studies in this guidance refer only to BA studies for ensuring
62 	product quality.
63
64 	This guidance should be used with other, more general CMC and BA and BE guidances available
65 	from CDER.3 Product quality information is different from, yet complementary to, the clinical
66 	safety and efficacy information that supports approval of an NDA. For information on the type of
67 	safety and efficacy studies that may be requested for a new active ingredient/active moiety
68 	intended for local action in the nose, or for a new product such as a nasal aerosol that may include
69 	an active ingredient/active moiety previously approved in a nasal spray, we recommend
70 	appropriate CDER review staff be consulted.
71
72 	Note: Detailed CMC information relevant to nasal aerosols and nasals sprays is presented in the
73 	final guidance Nasal Spray and Inhalation Solution, Suspension, and Spray Drug Products
74 	Chemistry, Manufacturing, and Controls Documentation.4 The document provides
75 	complementary information on the BA/BE testing methods recommended in this guidance.
76
3 Guidances are available on the Internet at http://www.fda.gov/cder/guidance/index.htm.
4 A draft guidance, Metered Dose Inhaler (MDI) and Dry Powder Inhaler (DPI) Drug Products: Chemistry,
Manufacturing, and Controls Documentation, was issued in October 1998. Once finalized, it will represent
the Agency’s thinking on this topic.
2

Contains Nonbinding Recommendations
Draft — Not for Implementation
77 	A. 	BA and BE Data
78
79 	Bioavailability is defined at 21 CFR 320.1 as “the rate and extent to which the active ingredient or
80 	active moiety is absorbed from a drug product and becomes available at the site of action. For
81 	drug products that are not intended to be absorbed into the bloodstream, bioavailability may be
82 	assessed by measurements intended to reflect the rate and extent to which the active ingredient or
83 	active moiety becomes available at the site of action.” Bioequivalence is defined as “the absence
84 	of a significant difference in the rate and extent to which the active ingredient or active moiety in
85 	pharmaceutical equivalents or pharmaceutical alternatives becomes available at the site of drug
86 	action when administered at the same molar dose under similar conditions in an appropriately
87 	designed study.” BA and BE are closely related, and the same approach used to measure BA in an
88 	NDA can generally be followed in establishing BE for an NDA or ANDA. Although BA may be
89 	comparative, establishing BE specifically involves a comparison of the BA of one product with
90 	the BA of another product. BE is usually established using (1) a criterion to allow the
91 	comparison, based on means and/or variances for BA measures, (2) a confidence interval for the
92 	criterion, and (3) a BE limit (goalpost) for the criterion.
93
94 	BA and BE data must be provided in accordance with the regulations.5 BA and BE can be
95 	established using in vivo (pharmacokinetic (PK), pharmacodynamic (PD), or clinical) and in
96 	vitro studies, or, in certain cases, using in vitro studies alone.6 BA and BE assessments for
97 	locally acting nasal aerosols and sprays are complicated because delivery to the sites of action
98 	does not occur primarily after systemic absorption. Droplets and/or drug particles are deposited
99 	topically. The drug is then absorbed and becomes available at local sites of action. A drug
100 	administered nasally and intended for local action has the potential to produce systemic activity,
101 	although plasma levels do not in general reflect the amount of drug reaching nasal sites of
102 	action. Systemic exposure following nasal administration can occur either from drug absorbed
103 	into the systemic circulation from the nasal mucosa, or after ingestion and absorption from the
104 	gastrointestinal tract (Daley-Yates et al., 2001). For these reasons, BA and BE studies generally
105 	would consider both local delivery and systemic exposure or systemic absorption.
106
107 	1. 	Local Delivery BA/BE Concepts
108
109 	For local delivery, BA is a function of several factors, including release of the drug
110 	substance from the drug product and availability to local sites of action. Release of the
111 	drug from the drug product produces droplet or drug particle sizes and distribution
112 	patterns within the nose that are dependent upon the drug substance, formulation, and
113 	device characteristics. Availability to local sites of action is usually a function of droplet
114 	or drug particle sizes and distribution patterns, as well as drug dissolution in the case of
115 	suspension products, absorption across mucosal barriers to nasal receptors, and rate of
116 	removal from the nose. From a product quality perspective, the critical issues are release
117 	of drug substance from drug product and delivery to the mucosa. Other factors are of
5 21 CFR 320.21, Requirements for submission of in vivo bioavailability and bioequivalence data.
6 21 CFR 320.24, Types of evidence to establish bioavailability or bioequivalence.
3

Contains Nonbinding Recommendations
Draft — Not for Implementation
118 	lesser importance.
119
120 	A critical question in assessing product quality BA and BE is the extent to which one can
121 	rely on in vitro methods alone, or upon in vitro methods plus clinical endpoints, to
122 	measure (benchmark) BA and/or establish BE. In vitro methods are less variable
123 	(Newman et al., 1995; Borgstrom et al., 1996; Suman et al., 2002), easier to control, and
124 	more likely to detect differences between products if they exist, but the clinical relevance
125 	of these tests, or the magnitude of the differences in the tests, can not always be clearly
126 	established. Clinical endpoints may be highly variable (Welch et al., 1991; Meltzer et al.,
127 	1998) and relatively insensitive to dose differences over an eightfold or higher dose range
128 	(Advisory Committee for Pharmaceutical Science, 2001), thus insensitive in detecting
129 	potential differences between products. However, clinical studies can unequivocally
130 	establish effectiveness of the drug product.
131
132 	In this guidance, the recommended approach for solution formulations of locally acting
133 	nasal drug products, both aerosols and sprays, is to rely on in vitro methods to assess BA.
134 	To establish BE, the recommended approach relies on (1) qualitative and quantitative
135 	sameness of formulation of test and reference products, (2) comparability in container
136 	and closure systems, and (3) in vitro methods that demonstrate equivalent performance.
137 	This approach is based on the premise that in vitro studies would be more sensitive
138 	indicators of drug delivery to nasal sites of action than would be clinical studies. For
139 	solution formulations, see Section IV.B.1.
140
141 	The recommended approach for establishing BA and BE of suspension formulations of
142 	locally acting nasal drug products, both aerosols and sprays, is to conduct in vivo studies
143 	in addition to in vitro studies.7 As with the solution formulation aerosols and sprays, to
144 	establish BE, the approach also relies on qualitative and quantitative sameness of
145 	formulation of test and reference products and comparability in container and closure
146 	systems. We recommend that in vitro studies be coupled with a clinical study for BA, or
147 	a BE study with a clinical endpoint (Section VI), to determine the delivery of drug
148 	substance to nasal sites of action. In vivo studies are recommended because of an
149 	inability at the present time to adequately characterize drug particle size distribution
150 	(PSD) in aerosols and sprays (Sections V.B.3, 4). Drug PSD in suspension formulations
151 	has the potential to influence the rate and extent of drug availability to nasal sites of
152 	action and to the systemic circulation.
153
154 	2. 	Systemic Exposure and Systemic Absorption BA/BE Concepts
155
156 	Locally acting drugs are intended to produce their effects upon delivery to nasal sites of
157 	action without relying on systemic absorption. Although systemic absorption may
158 	contribute to clinical efficacy for certain corticosteroids and antihistamines, the
7 Types of in vivo BE studies that may be submitted in support of an ANDA include, in addition to pharmacokinetic
studies, tests in humans in which an acute pharmacological effect is measured as a function of time and appropriately
designed comparative clinical trials for demonstration of BE (21 CFR 320.24).
4

Contains Nonbinding Recommendations
Draft — Not for Implementation
159 	consequences of systemic absorption (e.g., hypothalamic-pituitary-adrenal (HPA) axis
160 	suppression by corticosteroids) are generally undesirable. In the absence of validated in
161 	vitro methodology for characterizing drug PSD for suspension products and when
162 	measurable plasma levels can be obtained, this guidance recommends PK studies to
163 	measure systemic exposure BA or to establish systemic exposure BE (see Section VII).
164 	For suspension products that do not produce sufficient plasma concentrations to allow
165 	assessment of systemic exposure, clinical studies or BE studies with a pharmacodynamic
166 	or clinical endpoint are recommended to measure systemic absorption BA and establish
167 	systemic absorption BE, respectively (Section VIII). For a schematic representation of
168 	recommended studies, see Appendix A: Decision Tree.
169
170 	B. 	CMC and In Vitro BA Tests (Noncomparative) Versus BE Tests
171 	(Comparative)
172
173 	Generally, CMC tests help characterize the identity, strength, quality, purity, and potency of the
174 	drug product and assist in setting specifications (tests, methods, acceptance criteria) to allow
175 	batch release. These tests have a different purpose than do BA/BE tests, which focus on the
176 	release of the drug substance from the drug product. Some of the in vitro BA/BE tests described
177 	in this guidance may be the same as CMC tests for characterization and/or batch release. CMC
178 	and in vitro BA tests have acceptance criteria. In vitro BE tests have BE limits. A specification
179 	(test, method, acceptance criterion) for a CMC test for batch release or an in vitro BA test is
180 	usually based on general or specific manufacturing experience. For example, a CMC test such as
181 	dose content uniformity has acceptance criteria based on repeated manufacturing of batches. In
182 	contrast, BE tests have limits that are not usually based on manufacturing experience, but are
183 	part of equivalence comparisons between test and reference products. BE limits may be based
184 	on a priori judgments and may be scaled to the variability of the reference product (see
185 	Appendices C, E). When conducted premarket for an NDA, some of the in vitro BA tests
186 	described in this guidance can be noncomparative and serve primarily to document (benchmark)
187 	the product quality BA of a pioneer product.
188
189
190 	III. 	FORMULATION AND CONTAINER AND CLOSURE SYSTEM
191
192 	A. 	Formulation
193
194 	Particle size, morphic form, and state of solvation of an active ingredient have the potential to
195 	affect the BA of a drug product as a result of different solubilities and/or rates of dissolution.
196 	We recommend for an ANDA of a suspension formulation, data demonstrating comparable PSD
197 	and morphic form of the drug particles, size and number of drug aggregates in the dosage form,
198 	and hydrous or solvate form of the active drug in the dosage form to the reference listed drug, be
199 	provided, where possible. Where impossible, the rationale for not providing this full set of
200 	comparative data is requested. For suspension formulations marketed in more than one strength,
201 	we recommend that the drug substance in each strength product be micronized under identical
202 	parameters, and the PSD of the resultant bulk drug used in each product strength be identical.
203
5

Contains Nonbinding Recommendations
Draft — Not for Implementation
204 	B. 	Container and Closure System
205
206 	Nasal aerosols usually consist of the formulation, container, valve, actuator, dust cap, associated
207 	accessories, and protective packaging, which together constitute the drug product. Similarly,
208 	nasal sprays usually consist of the formulation, container, pump, actuator, protection cap, and
209 	protective packaging, which together constitute the drug product.
210
211 	For nasal aerosols and nasal sprays approved under an ANDA, we recommend BE be
212 	documented on the basis of validated in vitro and vivo tests, or, in the case of solutions, validated
213 	in vitro tests alone may be appropriate. Assurance of equivalence on the basis of in vitro tests is
214 	greatest when the test product uses the same brand and model of devices (particularly the
215 	metering valve or pump and the actuator) as used in the reference product. If this is infeasible,
216 	we recommend that valve, pump, and actuator designs be as close as possible in all critical
217 	dimensions to those of the reference product. We recommend that metering chamber volumes and
218 	actuator orifice diameters be the same. For a nasal spray, spray characteristics can be affected by
219 	features of the pump design, including the precompression mechanism, actuator design, including
220 	specific geometry of the orifice (Kublic and Vidgren 1998), and the design of the swirl chamber.
221 	The external dimensions of the test actuator are expected to ensure comparable depth of nasal
222 	insertion to the reference actuator. A test product is expected to attain prime within the labeled
223 	number of actuations for the reference product. We recommend you consider the volume of
224 	components of the device that must be filled to deliver an actuation, including the internal
225 	diameter and length of the diptube because this volume can influence the number of actuations
226 	required to prime a spray pump.
227
228
229 	IV. 	DOCUMENTATION OF BA AND BE
230
231 	A. 	NDAs
232
233 	For product quality, we recommend that in vitro BA studies be provided in NDAs for solution
234 	and suspension products, and in vivo BA studies be provided for suspension products. These
235 	data are useful as a benchmark to characterize the in vitro performance, and for suspensions, the
236 	in vivo performance of the product. Where the formulation and/or method of manufacture of the
237 	pivotal clinical trial product changes in terms of physicochemical characteristics of the drug
238 	substance, the excipients, or the device characteristics, BE data using in vitro tests (for solution
239 	and suspension products) and in vivo tests (for suspension products) may be useful in certain
240 	circumstances to ensure that the to-be-marketed product (T) is comparable to very similar
241 	clinical trial batches and/or to batches used for stability testing (R) (Section V.A.1). We
242 	recommend sponsors discuss the usefulness of these BE approaches with the appropriate CDER
243 	review staff.
244
245 	B. 	ANDAs
246
247 	For product equivalency, we recommend that the drug concentration in the test and reference
248 	product formulations not differ by more than ±5 percent. In addition, we recommend that the
6

Contains Nonbinding Recommendations
Draft — Not for Implementation
249 	inactive ingredients in the test product formulation be qualitatively (Q1 )8 the same and
250 	quantitatively (Q2 ) essentially the same as the inactive ingredients in the formulation of the
251 	reference listed drug, and the container and closure recommendations of Section III be followed.
252 	Quantitatively essentially the same has been determined by CDER to mean that the
253 	concentration or amount of the inactive ingredient(s) in the test product would not differ by more
254 	than ±5 percent of the concentration or amount in the reference listed drug. We recommend a
255 	side-by-side Q1 and Q2 comparison of the compositions of the test and reference listed drug
256 	formulations be provided. Please also provide a side-by-side comparison of the components of
257 	the container and closure system, listing brand and model, dimensions of critical components
258 	(Section IIIB), and engineering drawings if possible.
259
260 	1. 	Solution Formulations
261
262 	We believe in vitro tests alone can be relied on to document BE for nasal solution
263 	formulation products intended for local action. This approach is based on an
264 	understanding that for solution products, equivalent in vitro performance and adherence
265 	to Q1 and Q2 recommendations and to container and closure recommendations will
266 	ensure comparable delivery to the nasal mucosa and to the respiratory and
267 	gastrointestinal tracts. Suggested methodology and validation approaches for the
268 	recommended tests are provided in Section V. Suggested statistical methods to allow
269 	comparisons will be discussed in the appendices to this document. When in vitro data
270 	fail to meet acceptance criteria, the applicant is encouraged to modify the test product to
271 	attain equivalent in vitro performance. Because of insensitivity to potential differences
272 	between T and R, in vivo studies would not be sufficient in the face of failed in vitro
273 	studies.
274
275 	2. 	Suspension Formulations with PK Systemic Exposure Data
276
277 	To document BE for suspension formulation products intended for local action, we
278 	recommend both in vitro and in vivo data be used. In vivo studies would include both a
279 	BE study with a clinical endpoint (local delivery) and a pharmacokinetic study (systemic
280 	exposure). This approach is only applicable for those suspension formulation products
281 	that produce sufficiently high plasma concentrations of the moiety(ies) to be measured to
282 	allow reliable analytical measurement for an adequate length of time after nasal
283 	administration. Suggested methodology and validation approaches for the recommended
284 	tests are provided for in vitro studies in Section V, and for in vivo studies in Sections VI
285 	and VII. As with solutions, in vivo studies would not be sufficient in the face of failed in
286 	vitro studies (i.e., in vitro BE studies that fail to meet the statistical tests) even though the
287 	BE study with a clinical endpoint or the PK study meets the statistical test. Conversely,
288 	ANDAs with acceptable in vitro data, but with in vivo data that fail to meet the statistical
289 	tests, would be insufficient to establish BE.
290
291 	3. 	Suspension Formulations without PK Systemic Exposure Data
8 See 21 CFR 314.94(a)(9)(v).
7

Contains Nonbinding Recommendations
Draft — Not for Implementation
292
293 	For those products intended for local action that produce blood or plasma levels that are
294 	too low for adequate measurement, given current assay constraints, a BE study with a
295 	clinical endpoint to establish equivalent local delivery to nasal sites (Section VI) and a
296 	study with a pharmacodynamic or clinical endpoint to establish equivalent systemic
297 	absorption (Section VIII) are recommended. In vivo studies that meet the statistical test
298 	would not be sufficient in the face of in vitro studies that fail to document BE. As for
299 	suspensions with PK data, ANDAs with acceptable in vitro data, but with in vivo data
300 	that fail to meet the statistical tests, would be insufficient to establish BE.
301
302 	C. 	Postapproval Change
303
304 	This document does not cover postapproval changes. Sponsors planning such changes can
305 	consult the guidance for industry Changes to an Approved NDA or ANDA and contact the
306 	appropriate review division prior to instituting the change.
307
308
309 	V. 	IN VITRO STUDIES
310
311 	A. 	Batches and Drug Product Sample Collection
312
313 	1. 	NDAs
314
315 	We recommend in vitro BA studies for nasal aerosols and sprays be performed on
316 	samp les from three or more batches: a pivotal clinical trial batch to provide linkage of in
317 	vitro performance to in vivo data; a primary stability batch; and if feasible, a production-
318 	scale batch. This selection of batches will ensure consistency of in vitro performance
319 	among the three types of batches. If a production-scale batch is unavailable, a second
320 	pivotal clinical trial batch or second primary stability batch can be substituted. When
321 	three batches are studied, we recommend the batches be manufactured, preferably from
322 	three different batches of the drug substance, different batches of critical excipients, and
323 	different batches of container and closure components. However, the container (canister
324 	or bottle) can be from the same batch. We prefer that the three batches be studied at the
325 	same time, if possible, to remove interstudy variation from the estimation of between
326 	batch means and variances.
327
328 	The BA batches to be studied would be equivalent to the to-be-marketed product and
329 	representative of production scale. The manufacturing process for these batches would
330 	simulate that of large-scale production batches for marketing (additional information on
331 	large-scale batches is provided in the International Conference on Harmonisation (ICH)
332 	guidance for industry Q1A Stability Testing of New Drug Substances and Products,
333 	Section II.B.3). Complete batch records, including batch numbers of device
334 	components used in the batches, would accompany the BA submission.
335
336 	In vitro BA studies are intended to characterize the means and variances of measures of
337 	interest for canisters (nasal aerosols) or bottles (nasal sprays) within a batch and between
8

Contains Nonbinding Recommendations
Draft — Not for Implementation
338 	batches, where applicable. However, under 21 CFR 320.1 and 320.21, the studies can be
339 	noncomparative to other formulations or products. The in vitro tests and metrics are
340 	described in Section V.B of this guidance. The recommended number of canisters or
341 	bottles of each batch to be used in the above studies, and recommendations for statistical
342 	analyses, are described in Appendix B.
343
344 	2. 	ANDAs
345
346 	In vitro BE studies for nasal aerosols and sprays would generally be performed on
347 	samples from each of three or more batches of the test product and three or more batches
348 	of the reference listed drug. Test product samples would be from the primary stability
349 	batches used to establish the expiration dating period. When three batches are studied,
350 	we recommend the test product be manufactured, preferably from three different batches
351 	of the drug substance, different batches of critical excipients, and different batches of
352 	container and closure components. However, the container (canister or bottle) can be
353 	from the same batch. For nasal sprays formulated as solutions, in vitro BE tests can
354 	alternatively be performed on three sublots of product prepared from one batch of the
355 	solution.9
356
357 	The BE batches to be studied would be equivalent to the to-be-marketed product. The
358 	manufacturing process of these batches would simulate that of large-scale production
359 	batches for marketing. Complete batch records, including batch numbers of device
360 	components used in the batches or sublots (for solution nasal sprays) would accompany
361 	the BE submission.
362
363 	Reference product samples would be from three different batches available in the
364 	marketplace. The recommended in vitro tests and metrics are described in Section V.B.
365 	The recommended number of canisters or bottles of each product and batch to be used in
366 	the above studies, and recommended statistical approaches, are described in Appendices
367 	C, D and E.
368
369 	B. 	Tests and Metrics
370
371 	In vitro BA and BE for locally acting drugs delivered by nasal aerosol or nasal spray are usually
372 	characterized using seven tests:
373
374 	1. 	Single Actuation Content Through Container Life
375 	2. 	Droplet Size Distribution by Laser Diffraction
376 	3. 	Drug in Small Particles/Droplets, or Particle/Droplet Size Distribution by Cascade
377 	Impactor
378 	4. 	Drug Particle Size Distribution by Microscopy
9 For solution formulation nasal sprays, variability in in vitro BE study data between batches is expected to be due
primarily to variability in the device components of the product rather than in the solution. Therefore, a single batch
of solution can be split-filled into three equal size sublots of product. The sublots would be prepared from three
different batches of the same device (pump and actuator) components.
9

Contains Nonbinding Recommendations
Draft — Not for Implementation
379 	5. 	Spray Pattern
380 	6. 	Plume Geometry
381 	7. 	Priming and Repriming
382
383 	These tests are relevant to all nasal aerosols and nasal sprays, whether formulated as solution or
384 	suspension products, with the exception of drug particle size distribution by microscopy, which
385 	applies only to suspension products. The in vitro tests are summarized in Table 1.
386
387 	We recommend you validate all in vitro tests for accuracy and precision prior to the study. For
388 	applicable studies, instrument settings established during prestudy validation would be used in
389 	the study. For comparative studies, use of the same settings will ensure that T and R are studied
390 	under the same instrumental conditions. The in vitro tests would be conducted on canisters or
391 	bottles selected in a random manner from the test batch, including units from the beginning,
392 	middle, and end of the production run. Actuation should be conducted in a manner that removes
393 	potential operator bias, either by employing automatic actuation, or by employing blinded
394 	procedures when manual actuation is used. However, we recommend automated actuation
395 	systems for all comparative in vitro BE tests. These systems are expected to decrease variability
396 	in drug delivery due to operator factors, thereby increasing the sensitivity for detecting potential
397 	differences between products in the above tests.10 In addition, it is important that the analyst
398 	performing the postactuation evaluations of the collected data be blinded to the identity of the
399 	samples. We recommend analytical methods used for analysis of samples from the in vitro tests
400 	be validated.11 Unexpected results and deviations from protocol or SOPs, with justification for
401 	deviations, would be reported. Examples include, but are not limited to, canisters or bottles
402 	replaced during in vitro analyses, failure to use the specific actuations required by the protocol,
403 	and experiments rejected due to assignable causes (e.g., instrument failure, sample collection, or
404 	processing errors). The original and reanalyzed data, with the reason for reanalysis, would be
405 	tabulated in the study report. The validation reports for the in vitro tests and analytical methods,
406 	the randomization procedure, and all test methods or SOPs for each test would accompany the
407 	data in the submission. When appropriate, we recommend the test method or SOP include a
408 	standardized shaking procedure prior to testing, following labeled instructions, if any.
409
410 	In addition to submission of all raw data, the agency would like to see supporting documentation
411 	for the following tests: Droplet Size Distribution by Laser Diffraction, Spray Pattern, and Plume
412 	Geometry. Documentation includes instrument output reports and photographic or graphic
413 	material as applicable. We recommend that documents be clearly labeled to indicate the product
414 	(e.g., T or R), batch number, and testing conditions (e.g., distance, lifestage, delay time), as
10 Automatic actuation systems can be stand-alone or accessories for spray characterization instruments.
Systems can include settings for force, velocity, acceleration, length of stroke, and other relevant parameters.
Selection of appropriate settings would be relevant to proper usage of the product by the trained patient, and for
nasal sprays, may be available from pump suppliers for tests such as Droplet Size Distribution by Laser
Diffraction and Spray Pattern. In the absence of recommendations from the pump supplier, we recommend that
settings should be documented based on exploratory studies in which the relevant parameters are varied to
simulate in vitro performance upon hand actuation. Selected settings used for the in vitro studies would be
specified in the test method or SOP for each test for which the system is employed.
11 A draft guidance for industry entitled Analytical Procedures and Methods Validation was issued in August 2000.
10

Contains Nonbinding Recommendations
Draft — Not for Implementation
415 	appropriate. For Droplet Size Distribution by Laser Diffraction, profiles of droplet size and
416 	obscuration or percent transmission over the complete life of the single sprays would be
417 	submitted. For Spray Pattern and Plume Geometry, we recommend each image display the
418 	relevant BA/BE measures described in this guidance. Supporting documentation for Droplet
419 	Size Distribution by Laser Diffraction, Spray Pattern, and Plume Geometry would include
420 	representative copies, preferably electronic, of >20 percent of the total observations. For Spray
421 	Pattern and Plume Geometry quantitated by automatic image analysis, representative electronic
422 	images rather than paper copies of >20 percent of the total observations would be submitted, as
423 	electronic files are definitive. 	For automated image analysis of Spray Pattern and Plume
424 	Geometry, in addition to the electronic images, we recommend paper copies of a few screen
425 	images be submitted as reference samples.
426
427 	1. 	Single Actuation Content (SAC) Through Container Life
428
429 	For noncomparative data, SAC through container life testing is used to characterize the
430 	delivery of drug discharged from the actuator of an aerosol or nasal spray relative to label
431 	claim through container life. For comparisons of T and R products, this test ensures that
432 	the T product delivers an equivalent amount of drug relative to the R product over the
433 	labeled number of actuations. The tests are distinct from and do not apply dose content
434 	uniformity (DCU) or spray content uniformity (SCU) acceptance criteria.
435
436 	The dosage unit sampling apparatus for collection of an emitted dose from an aerosol is
437 	described in U.S. Pharmacopeia (USP) 25, <601>. We recommend a suitable apparatus
438 	be used for collecting an emitted dose from a nasal spray. For both solution and
439 	suspension formulations of nasal aerosols and nasal sprays, the mass of drug per
440 	actuation would be based on a stability-indicating chemical assay unless use of a
441 	nonstability-indicating method is justified. Because the data at beginning (B) lifestage
442 	will also be used for confirmation of priming (Section V.B.7), SAC through container life
443 	would be based on single actuation data per determination. 	For BA and BE
444 	submissions, the tests would determine delivered (emitted or ex-actuator) drug mass from
445 	primed units at the beginning of unit life, at the middle of unit life, and at the end of unit
446 	life12 for nasal aerosols, and at beginning and end of unit life for nasal sprays. The
447 	delivered mass of drug substance would be expressed both as the actual amount and as a
448 	percentage of label claim. We recommend that mean and variability in SAC through
449 	container life be determined based on within and between unit (container) data and
450 	between batch (or sublot) data. For BE data, equivalence of T and R data would be based
451 	on the statistical methodology of Appendix C.
452
453 	To use the SAC through container life data for priming studies, we recommend aerosols
454 	and sprays be unprimed prior to the conduct of the tests. Therefore, for aerosols, the test
12 Based on the labeled number of actuations, this guidance uses the terms beginning lifestage (B), middle lifestage
(M), and end lifestage (E) interchangeably with the terms beginning of unit life (the first actuation(s) following the
labeled number of priming actuations); middle of unit life (the actuation(s) corresponding to 50 percent of the labeled
number of actuations); and end of unit life (the actuation(s) corresponding to the label claim number of actuations).
11

Contains Nonbinding Recommendations
Draft — Not for Implementation
455 	would be performed at such time that the product meets two conditions: (1) after the
456 	lagering period and (2) not less than one month after the last actuation conducted as part
457 	of batch release testing. During the time period between batch release and SAC through
458 	container life testing, the aerosol product would not be actuated. Also, during this one
459 	month period, both T and R aerosols would be stored in the valve upright position, unless
460 	labeling indicates that the product be stored in the valve down position, in which case the
461 	test would be conducted on products stored in the valve down position. For sprays, the
462 	SAC through container life test would be conducted not less than one month after
463 	completion of batch release testing. During the time period between batch release and
464 	SAC testing, the product would not be actuated.
465
466 	2. 	Droplet Size Distribution by Laser Diffraction
467
468 	Droplet size distribution is an important property influencing the nasal deposition of
469 	aerosols and sprays, and we recommend that it be thoroughly characterized.
470
471 	a. 	Nasal sprays
472
473 	We recommend that droplet size distribution be determined using laser diffraction
474 	or an appropriately validated alternate methodology.
475
476 	Laser diffraction is a nonaerodynamic optical method of droplet sizing that
477 	measures the geometric size of droplets in flight. Modern laser diffraction
478 	instrumentation can provide plots of obscuration (optical concentration) or
479 	percent transmission (%T) and droplet size distribution (D10 , D50 , D90 ) over the
480 	entire life of a single spray. Span ((D90 - D10 )/D50 ) can be computed from these
481 	data. These profile data indicate that each plume can be characterized by three
482 	phases: formation, fully developed, and dissipation. For nasal sprays, the general
483 	profile for obscuration or percent T versus time can be characterized by a rapid
484 	increase in obscuration, or decrease in percent T, early in the life of the spray
485 	(formation phase), followed by attainment of a plateau (fully developed phase),
486 	then a rapid decrease in obscuration, or increase in percent T, late in the life of the
487 	spray (dissipation phase). Changes in droplet size occur coincident with the
488 	changes in obscuration or percent T, with droplet sizes attaining plateau values
489 	within the same approximate time period as the plateau in obscuration or percent
490 	T. Profiles of the droplet size and obscuration or percent T over the complete life
491 	of the single sprays are recommended to be determined at each of two distances
492 	(see below) to establish the fully developed phase during which data would be
493 	collected. Droplet size distribution and span during the fully developed phase are
494 	requested. The sponsor’s protocol or SOP would state the criterion selecting the
495 	region of the plateau at which droplet size data will be determined (e.g., the
496 	average of all scans over the entire plateau, the data of a single scan (sweep) only
497 	at the maximum obscuration (or minimum percent T), or the average of a
498 	specified range of scans around this obscuration or percent T). This criterion
499 	would be established prior to the study for each of the two distances and
500 	implemented consistently during the study.
12

Contains Nonbinding Recommendations
Draft — Not for Implementation
501
502 	We would also like to see instrument setup and operation conditions. We
503 	recommend the instrument be operated within the manufacturer’s recommended
504 	obscuration or percent T range, which would be stated in the submission, to
505 	avoid or minimize multiple scattering (due to high droplet concentration).
506 	Avoidance of multiple scattering is preferred to use of a correction algorithm that
507 	compensates for this effect.
508
509 	Single spray droplet size distribution and span would be reported based on
510 	volume (mass) rather than count (number of droplets). We would like to request
511 	data be provided for nasal sprays at:
512
513 	• 	Fully developed phase only
514 	• 	B and E lifestages
515 	• 	Two distances from the actuator orifice. For increased ability to detect
516 	potential differences between products, it is recommended that the studies be
517 	performed within a range of 2 to 7 cm from the orifice, with the two distances
518 	separated by 3 cm or more.
519
520 	b. 	Nasal aerosols
521
522 	Droplet size distribution can be determined using laser diffraction or
523 	appropriately validated alternate methodology.
524
525 	We would like to see instrument setup and operation conditions. We recommend
526 	the instrument be operated within the manufacturer’s recommended obscuration
527 	or percent T range, which would be stated in the submission, to avoid or
528 	minimize multiple scattering (due to high droplet concentration). Avoidance of
529 	multiple scattering is preferred to use of a correction algorithm that compensates
530 	for this effect.
531
532 	Beam steering resulting from refractive index effects due to evaporation of
533 	propellant is an additional concern for nasal aerosols. Droplet size distribution
534 	would be characterized at distances from the actuator that eliminate or minimize
535 	beam steering, if possible. If a correction algorithm is used, we recommend an
536 	explanation of the corrections be provided.
537
538 	We ask that single-spray droplet size distribution and span be reported based on
539 	volume (mass) rather than count (number of droplets). Data would be provided
540 	for nasal aerosols at:
541
542 	• 	Fully developed phase only
543 	• 	B and E lifestages
544 	• 	Two distances from the actuator orifice
545
546 	For both nasal sprays and nasal aerosols, mean D10 , D50 , D90 values for a given bottle or
13

Contains Nonbinding Recommendations
Draft — Not for Implementation
547 	canister can be computed from the mean of up to three consecutive sprays from that unit
548 	at each lifestage. However, to assess precision, the data of each spray would also be
549 	reported.
550
551 	3. 	Drug in Small Particles/Droplets, or Particle/Droplet Size Distribution by
552 	Cascade Impactor
553
554 	Sizing of droplets or particles by multistage cascade impactor (CI) measures
555 	aerodynamic diameter based on inertial impaction, an important factor in the
556 	deposition of drug in the nasal passages. Analytical data should be based on a
557 	validated chemical assay.11 We recommend that analytical runs include at least
558 	three or more concentrations of quality control samples that represent the entire
559 	range of the standard curve or the expected concentration range of samples from
560 	the various stages of the CI. An analytical validation report would accompany the
561 	CI data report. The SOP or validation report would indicate the minimum
562 	quantifiable mass of drug deposited on each location reported.
563
564 	a. 	Nasal Sprays: Drug in Small Particles/Droplets
565
566 	For nasal sprays, the majority of the emitted dose is deposited prior to or on the
567 	first stage of the CI test. Small droplets, for this test and dosage form defined as
568 	smaller in size than the nominal effective cutoff diameter (ECD) of the top stage
569 	of a suitable CI, may potentially be delivered to regions of the airways beyond
570 	the nose. This test is intended to determine the amount of drug in small
571 	particles/droplets. For example, for USP 25 Apparatus 1 (<601>), an eight
572 	stage CI operated with the standard 28.3 liter per minute configuration, small
573 	droplets are those under 9.0 microns. For BA, the CI test is intended to quantify
574 	the mass of drug in small droplets. For BE, the mass of drug in small droplets
575 	for the T product would be less than or equivalent to the corresponding mass of
576 	drug from the R product. The comparative test addresses a potential safety
577 	concern — an excess of small droplets due to T relative to R might deliver to
578 	regions beyond the nose excipients with possible adverse pulmonary effects.
579 	The CI test for nasal sprays is not intended to provide PSD of drug or
580 	aerosolized droplets.
581
582 	Measurable levels of drug below the top stage of the CI would be a function of the
583 	specific drug product and the experimental setup and procedure, including the
584 	number of actuations and assay sensitivity. Thus, we recommend a validated,
585 	highly sensitive assay be used. In Agency experience, a two-liter or larger
586 	induction port (expansion chamber) is preferred to a one-liter chamber. We prefer
587 	studies use the fewest number of actuations (generally not exceeding 10) justified
588 	by the sensitivity of the assay, to be more reflective of individual doses. Drug
589 	deposition would be reported in mass units. Mass balance accountability would
590 	be reported. Mass balance would be based on drug deposition on each of
591 	valvestem, actuator, adapters, induction port, any other accessories, the top stage,
592 	and all lower stages to the filter. The total mass of drug collected on all stages
14

Contains Nonbinding Recommendations
Draft — Not for Implementation
593 	and accessories is recommended to be between 85 and 115 percent of label claim
594 	on a per actuation basis. The total mass of drug below the top stage is of primary
595 	interest. Therefore the pooled mass of drug deposited on all lower stages and
596 	filter can be reported.
597
598 	For BA and BE, CI test would be data requested only at the beginning lifestage.
599 	Statistical approaches will be provided in Appendices B and D, respectively.
600
601 	b. 	Nasal Aerosols: Particle/Droplet Size Distribution
602
603 	CI studies for nasal aerosols would use an induction port (expansion chamber)
604 	that maximizes drug deposition below the top stage of the CI. For this reason, a
605 	one-liter induction port is preferred to the USP 25 (<601>) induction port,
606 	although other sizes may also be appropriate. Agency experience indicates that
607 	with a suitable induction port and CI, the amount of drug deposited below the top
608 	stage from nasal aerosols formulated with either chlorofluorocarbon or
609 	hydrofluoroalkane propellants is of the same order of magnitude as from orally
610 	inhaled aerosols. Therefore, unlike for nasal sprays in which the total mass of
611 	drug below the top stage is of interest, we recommend a particle/droplet size
612 	distribution be provided for this dosage form. Selection of the most suitable CI
613 	may be influenced by the effective cutoff diameters (ECDs) of stages of various
614 	brands of cascade impactors, the geometry of the induction port, and other factors.
615 	The number of actuations recommended for the CI study of aerosols is described
616 	in the draft guidance Metered Dose Inhaler (MDI) and Dry Powder Inhaler
617 	(DPI) Drug Products: Chemistry, Manufacturing, and Controls Documentation.
618 	Drug deposition would be reported in mass units. Mass balance accountability
619 	would be reported.
620
621 	For BA and BE, CI data would be requested only at the beginning lifestage. At
622 	this time, it is recommended that studies of nasal aerosols use USP 25 Apparatus
623 	1 (<601>) operated at the standard 28.3 liter per minute configuration. We
624 	recommend determination of a profile based on drug deposition at 11 sites: (1)
625 	sum of valve stem plus actuator; (2) induction port; (3 - 10) eight individual
626 	stages; and (11) filter. Deposition in the valve stem plus actuator would be
627 	included to provide a profile of drug deposition ex-valve rather than ex-actuator.
628 	It should be noted that the in vitro BE limit for the profile comparison depends on
629 	the number of stages and other accessory deposition sites. Statistical approaches
630 	for BA and BE will be provided in Appendices B and E, respectively.
631
632 	4. 	Drug Particle Size Distribution by Microscopy
633
634 	For suspension products, drug particle size may be important for rate of
635 	dissolution and availability to sites of action within the nose. Therefore, drug
636 	particle size distribution (PSD) and extent of agglomerates would be
637 	characterized in the spray or aerosol formulation prior to actuation, and in the
638 	spray following actuation. Determination of PSD and agglomerates in both the
15

Contains Nonbinding Recommendations
Draft — Not for Implementation
639 	formulation and following actuation are intended to characterize the potential
640 	influence of the device on deagglomeration. Determination in the spray is only
641 	requested at the beginning lifestage. Nasal spray formulations frequently contain
642 	suspended drug substance in the presence of insoluble suspending agent, which
643 	complicates the particle size characterization. When examining formulations
644 	containing suspending agents, and currently available technology cannot be
645 	acceptably validated to determine drug particle size, a qualitative and semi
646 	quantitative method for examination of drug and aggregated drug particle size
647 	distribution can be used. We recommend studies of nasal sprays include placebo
648 	product to provide an estimate of the occurrence of apparent drug particles (false
649 	positives) due to excipient. Evaluation may use light microscopy or other
650 	appropriate means.
651
652 	For NDAs and ANDAs of both sprays and aerosols, we recommend drug PSD
653 	and agglomerates data be provided in the BA or BE submission, along with a
654 	description of the test method. Sponsors can submit representative
655 	photomicrographs, if desired. For BE, PSD by light microscopy, even if
656 	qualitative or semi-quantitative, can be useful to the applicant to estimate
657 	particle size relative to the precursor product prior to further product
658 	development and testing. These data are supportive, and formal statistical
659 	testing is not applicable.
660
661 	5. 	Spray Pattern
662
663 	Spray pattern studies characterize the spray either during the spray prior to
664 	impaction, or following impaction on an appropriate target such as a thin-layer
665 	chromatography (TLC) plate. Spray patterns for certain nasal spray products
666 	may be spoked or otherwise irregular in shape.
667
668 	Spray patterns can be characterized and quantitated by either manual or
669 	automated image analysis, if validated. Both analyses will allow shape and size
670 	to be determined. Automated analysis systems may also allow determination of
671 	center of mass (COM; unweighted for image intensity) and/or center of gravity
672 	(COG; weighted for image intensity) within the pattern to be determined. COG
673 	is of greater interest and is preferred in the automated analyses of spray patterns.
674 	Automated image analysis is expected to increase objectivity in spray pattern
675 	measurement. The technology enables the perimeter of the true shape of the
676 	spray pattern to be determined, identifies COM and/or COG, and enables the area
677 	within the perimeter to be quantitated, thus its use is encouraged.
678
679 	Equivalence of spray patterns between T and R products can be established
680 	based on a combination of qualitative and quantitative measures:
681
682 	• 	Comparative visual inspection for shape. For the automated analyses, the true
683 	shapes identified by the software serve as the basis of comparison
684 	(qualitative). Establishment of qualitative sameness of T and R spray pattern
16

Contains Nonbinding Recommendations
Draft — Not for Implementation
685 	shapes is a prerequisite to the quantitative analyses in the following two
686 	bullets.
687 	• 	Equivalent area within the perimeter of the true shape for automated
688 	analysis, or equivalent Dmax for manual analysis (quantitative).
689 	• 	Equivalent ovality (ellipticity) ratio (quantitative).
690
691 	a. 	For nonimpaction systems
692
693 	Spray patterns can be visualized using a system based on a laser light sheet and
694 	high-speed digital camera that enables visualization of a pattern perpendicular to
695 	the axis of the nasal spray. The perimeter of the true shape, area within the
696 	perimeter (to include a high proportion, e.g., >95 percent of the total pattern),
697 	COG, and Dmax (longest diameter) and D min (shortest diameter) that pass through
698 	the COG and extend to the perimeter of the true shape, can be determined based
699 	on automated analysis using time-averaged images over the duration of a single
700 	spray. Software settings can be established during prestudy validation and the
701 	settings should be used consistently in the study. Statistical analysis at each
702 	distance would be based on equivalence of area within the perimeter and ovality
703 	ratio (Dmax divided by Dmin ).
704
705 	b. 	For impaction systems
706
707 	The number of sprays per spray pattern would preferably be one. We recommend
708 	that the visualization technique be specific for the drug substance. If exploratory
709 	studies document that a drug-specific reagent cannot be found, a nonspecific
710 	visualization reagent can be used. We recommend that application of the reagent
711 	be controlled to maintain the details of the image intensity of the pattern.
712
713 	Manual analysis
714
715 	The approximate COM would be identified, and Dmax and D min drawn through this
716 	center. The two lines may not be orthogonal to each other. Representative plots
717 	can be submitted, and each figure can be marked with the COM, Dmax and D min ,
718 	each based on visual analysis. The ovality ratio would be provided for each spray
719 	pattern. Statistical analysis at each distance would be based on equivalence of
720 	Dmax and ovality ratio.
721
722 	Automated analysis
723
724 	The automated image analysis software can define the perimeter of the true shape
725 	of the spray pattern to include a high proportion (e.g., >95 percent) of the total
726 	pattern. T and R would both be sprayed on each TLC plate to ensure
727 	measurement of the spray pattern at the same intensity range for a given plate.
728 	Dmax and D min would pass through the COM or the COG, as appropriate, and
729 	extend to the perimeter of the true shape. Statistical analysis at each distance
730 	would be based on equivalence of area within the perimeter and ovality ratio.
17

Contains Nonbinding Recommendations
Draft — Not for Implementation
731
732 	c. 	For both nonimpaction and impaction systems
733 	The information above would apply to spray patterns in which the COM or COG
734 	falls within the perimeter of the image of the actual spray pattern, and the Dmax
735 	axis doesn’t extend outside of the perimeter. Infrequently, the COM or COG may
736 	fall outside the perimeter of the spray pattern, and/or the Dmax axis may cross the
737 	perimeter. Horseshoe-shaped and certain other patterns may cause such an effect.
738 	When this occurs, automated analysis using a system that has the capability of
739 	fitting the perimeter with an appropriate geometric shape is recommended.
740 	Statistical analysis at each distance would be based on equivalence of area within
741 	the perimeter of the true shape of the spray pattern (not within the fitted
742 	geometric shape), and ovality ratio, where D max and D min are computed from the
743 	fitted geometric shape (e.g., ellipse).
744
745 	For all cases above, we recommend spray patterns be determined based on:
746
747 	• 	Single actuations (nonimpaction systems), or preferably single actuations
748 	(impaction systems)
749 	• 	Beginning lifestage only
750 	• 	Two distances from the actuator orifice, which allow discriminatory
751 	capability between individual pump units and between T and R products.
752 	For nasal sprays, these distances are recommended to be at least 3 cm
753 	apart within the range of 3 to 7 cm.
754
755 	For manual quantitation of spray patterns based on impaction studies such as TLC
756 	plate methodology, we recommend the submission include copies, preferably
757 	electronic, of images of representative spray patterns at two distances, and each
758 	figure would clearly indicate the estimated COM (manual analysis), Dmax and
759 	D min . When automated image analysis software is used for impaction studies, data
760 	would be presented in electronic files. For automated image analysis of either
761 	impaction or nonimpaction studies, electronic files would be definitive.
762 	Submission of electronic files is recommended to avoid printer-dependent
763 	variations in spatial calibration of images. These files would contain the images,
764 	showing the COG or COM and the perimeter of the true shape of the spray
765 	pattern, and the accompanying quantitation reports. Each image would also
766 	include a legible scale used for measurement.
767
768 	Some automated image analysis software may not include automated quantitation
769 	of spray pattern images. For such cases, the analyst would determine and display
770 	the quantitative parameters on the electronic image. As mentioned above,
771 	quantitation of electronic images would be definitive.
772
773 	6. 	Plume Geometry
774
775 	Plume geometry describes a side view of the aerosol cloud parallel to the axis of
18

Contains Nonbinding Recommendations
Draft — Not for Implementation
776 	the plume, and we recommend it be based on high-speed photography, a laser
777 	light sheet and high speed digital camera, or other suitable methods. The image
778 	would be snapshot, not time-averaged. Quantitation can be by manual analysis or
779 	automated image analysis.
780
781 	During the very early life of an aqueous nasal spray plume, formulation may exit
782 	the actuator orifice as a narrow stream that subsequently forms a relatively stable,
783 	fully developed, conical plume prior to separating from the orifice. We
784 	recommend plume angle, width, and height, all quantitated by the same analytical
785 	method, be reported at a single delay time while the fully developed phase of the
786 	plume is still in contact with the actuator tip. The applicant would provide
787 	documentation that the plume is fully developed at the selected delay time. The
788 	angle would be based on the conical region of the plume extending from a vertex
789 	that occurs at or near the actuator tip. Plume angle based on spray pattern
790 	dimensions and distance from actuator tip to an impaction surface is not
791 	appropriate. For this guidance, the recommended plume width would be the
792 	width at a distance equal to the greater of the two distances selected for
793 	characterization of the spray pattern. Plume width data would thus be
794 	complementary to spray pattern data obtained at the same distance. Plume height
795 	would be the distance from the actuator orifice (sprays) or end of the inhaler tube
796 	(aerosols) to the leading edge of the plume. We request that the criteria for
797 	defining the plume angle, width, and height borders be provided.
798
799 	Plume geometry would be performed at:
800
801 	• 	Beginning lifestage only
802 	• 	One side view only
803 	• 	A single delay time
804
805 	The submission would include photographs when quantitation is by manual
806 	analysis, or digital images when quantitation is by automated image analysis.
807 	Each image would also include a legible scale used for measurement, and the
808 	delay time would be clearly indicated. Images would clearly indicate the plume
809 	angle, width, and height. When automated image analysis is used, quantitation of
810 	electronic images would be definitive. Manual quantitation based on paper copies
811 	of electronic images would not be appropriate.
812
813 	We recommend plume geometry measurements be summarized as mean,
814 	geometric mean, and %CV. Comparative data would be supportive, thus for BE
815 	studies, the ratio of the geometric mean of the three batches of T to that of the
816 	three batches of R, based on log transformed data, would fall within 90 – 111
817 	percent (point estimates) for plume angle and width. Due to subjectivity in the
818 	measurement of plume height, point estimates would not be applicable.
819
820 	7. 	Priming and Repriming
19

Contains Nonbinding Recommendations
Draft — Not for Implementation
821
822 	Priming and repriming data will ensure delivery of the labeled dose of drug
823 	following labeled instructions for use. Priming would be established based on the
824 	same B lifestage data obtained for the single actuation content (SAC) through
825 	container life study (Section V.B.1). For products approved under an NDA,
826 	priming and repriming data based on single actuations would be provided in the
827 	CMC portion of the submission.
828
829 	For products approved under an ANDA, the labeling would be the same as that for
830 	the R product, except for specific changes described in the regulations (21 CFR
831 	314.94(a)(8)(iv)). For nasal sprays and some nasal aerosols, the R product
832 	labeling (package insert and/or patient package insert) describes the number of
833 	actuations to prime the product on initial use and on repriming following one or
834 	more periods of nonuse (e.g., 24 hours and 7 days following last dose). For these
835 	products, we request priming and repriming data for T and R products. Studies
836 	would follow the recommended time periods described in Section V.B.1 between
837 	lagering and/or batch release testing and conduct of the priming test. Priming
838 	and/or repriming studies would not be requested when the R product lacks priming
839 	and/or repriming instructions, respectively.
840
841 	We recommend that priming and repriming data for T in multiple orientations
842 	be provided in the CMC portion of the ANDA submission. Therefore, for the
843 	BE submission, studies can be based on products stored in the valve upright
844 	position, with the exception of nasal aerosols in which R labeling recommends
845 	storage in the valve down position. For the latter products, priming data, and
846 	repriming data when applicable, would be provided following storage in the
847 	valve down position. Priming studies would be based on the emitted dose of the
848 	single actuation at B lifestage immediately following the specified number of
849 	priming actuations in the R product labeling. For ANDAs, priming would be
850 	established providing that the geometric mean emitted dose of the 30 canisters
851 	or bottles calculated from the SAC data at B lifestage falls within 95 – 105
852 	percent of label claim. Repriming would be similarly established based on a
853 	single actuation following the specified number of repriming actuations in the R
854 	product labeling. Although noncomparative to R, the priming studies would be
855 	essential to the BE submission to document that each product delivers the
856 	labeled dose within the number of actuations stated in the R product labeling,
857 	thus ensuring that the SAC through container life studies are conducted on
858 	primed T and R products.
859
860
861 	VI. 	CLINICAL STUDIES FOR LOCAL DELIVERY
862
863 	A. 	General Information
864
865 	1. 	NDAs
866
20

Contains Nonbinding Recommendations
Draft — Not for Implementation
867 	At the present time, of the classes of drugs covered in this guidance, only certain
868 	corticosteroids are formulated as suspension formulation nasal aerosols and nasal
869 	sprays and require in vivo studies as a component of the BE or BA submission
870 	(21 CFR 320.21). The same adequate and well-controlled clinical trials in humans
871 	conducted under an authorized IND, used to establish the safety and effectiveness
872 	of a drug product in support of a forthcoming NDA (21 CFR 314.126), can be
873 	used in some cases to establish BA or, when comparative, BE (21 CFR 320.24).
874
875 	2. 	ANDAs
876
877 	Clinical studies are at times incapable of showing a dose-response relationship
878 	and may not be consistently reproducible. However, a showing of dose
879 	response is not necessary for BE studies with a clinical endpoint, as these studies
880 	are intended only to confirm the lack of important clinical differences between T
881 	and R suspension formulation nasal aerosol and nasal spray products (Advisory
882 	Committee for Pharmaceutical Science, 2001). 	For an ANDA, an authorized
883 	Bio-IND will be needed for the conduct of a BE study with a clinical endpoint.13
884
885 	A determination of bioequivalence of rhinitis BE study with a clinical endpoint
886 	for locally acting nasal suspension drug products would be based on the
887 	following premises for T relative to R products:
888
889 	• 	Qualitative and quantitative sameness of formulation
890 	• 	Comparability in container and closure systems
891 	• 	Equivalence of in vitro tests
892 	• 	Equivalence of systemic exposure or systemic absorption
893 	• 	Equivalence of the local delivery study.
894
895 	A number of FDA guidances provide information about the general conduct of
896 	clinical studies, including clinical studies to document BA and BE: General
897 	Considerations for Clinical Trials (International Conference on Harmonisation
898 	(ICH) E8); Structure and Content of Clinical Study Reports (ICH E3); Good
899 	Clinical Practice: Consolidated Guidance (ICH E6); Statistical Principles for
900 	Clinical Trials (ICH E9), and Choice of Control Group and Related Issues in
901 	Clinical Trials (ICH E10).
902
903 	B. 	Clinical Study Batches
904
905 	We recommend that the batch used for the BA study be the same pivotal clinical trial batch used
906 	in the in vitro BA studies (Section V.A). Where BE studies are conducted for an NDA, the
907 	batches of test and reference products would be the same batches employed in the in vitro testing.
13 Office of Generic Drugs Policy and Procedure Guide # 36-92, Submission of an "Investigational New Drug
Application" to the Office of Generic Drugs (OGD), October 13, 1992.
21

Contains Nonbinding Recommendations
Draft — Not for Implementation
908 	Each of the T and R batches used to establish local delivery BE for an ANDA would be one of
909 	the three batches used for the in vitro BE studies. We recommend that the inactive ingredients of
910 	the placebo (P) product meet Q1 and Q2 recommendations relative to the R product (Section
911 	IV.B); the P container and closure would meet the recommendations of Section III.B.
912
913 	C. 	Clinical BE Study Design and Subject Inclusion Criteria
914
915 	The study design would be the traditional treatment study in which T and R are assessed for a
916 	two-week duration. The two-week duration, in addition to allowing a comparison of equivalent
917 	efficacy, will also allow for an assessment of safety and tolerability over a reasonable period of
918 	use. We recommend the study be conducted at the lowest labeled adult recommended dose in
919 	an attempt to optimize study sensitivity. Prime products according to labeling instructions prior
920 	to dosing. Ensure that priming occurs out of range of the patients, to avoid inhalation of drug
921 	fired to waste. Documentation would rely on the inclusion of a test product placebo (P) dosed
922 	at the same frequency and number of actuations per nostril as T and R.
923
924 	A study population consisting of seasonal allergic rhinitis (SAR) patients will allow
925 	documentation of BE, which may extend to all indications in product labeling for locally acting
926 	nasal corticosteroids. 	In addition to a history of SAR, we recommend patients have a positive
927 	test for relevant specific allergens (e.g., allergen skin test) and be experiencing a defined
928 	minimum level of symptom severity at the time of study enrollment. 	We discourage the
929 	inclusion of patients with other significant diseases including asthma, with the exception of mild
930 	intermittent asthma.
931
932 	The recommended design for this study is a randomized, double-blind, placebo-controlled,
933 	parallel group study of 14 days duration, preceded by a 7-day placebo run-in period to establish a
934 	baseline and to identify placebo responders.14 We recommend placebo responders be excluded
935 	from the study to increase the ability to show a significant difference between active and placebo
936 	treatments (efficacy analysis), and to increase sensitivity to detect potential differences between
937 	T and R products (equivalence analysis). 	The protocol would define placebo responders a
938 	priori. Whether the drug is labeled for once or twice daily dosing, clinical evaluations would be
939 	made twice daily (AM and PM, 12 hours apart at the same times daily) throughout the 7-day
940 	placebo run-in period and the 14-day randomized treatment period. Scoring should be made
941 	immediately prior to each dose, to reflect the previous 12 hours (reflective scores) and how the
942 	patient is feeling at the time of evaluation (instantaneous or snapshot scores). Because the
943 	primary BE endpoint would be based on reflective symptom scores, placebo responders should
944 	be identified based on reflective scores, although BE endpoints would include both reflective and
945 	instantaneous scores.
946
947 	We recommend baseline scoring preferably consist of reflective AM and PM scoring on Days 5,
948 	6, and 7 of the placebo run-in period, and AM scoring (prior to drug dosing) on Day 1 of the 14
949 	day randomized treatment period, resulting in 7 total AM and PM ratings. 	Placebo responders
950 	would be identified based on the mean total nasal symptom score (TNSS) over the 7 total AM and
14 A draft guidance for industry entitled Allergic Rhinitis: Clinical Development Programs for Drug Products was
issued in April 2000. This guidance discusses general protocol issues including blinding. Once finalized, it will
represent the Agency’s thinking on this topic.
22

Contains Nonbinding Recommendations
Draft — Not for Implementation
951 	PM ratings. 	The study protocol would state the minimum qualifying reflective TNSS for
952 	enrollment at screening, and the same minimum qualifying TNSS would be met based on the
953 	mean of the 7 total AM and PM ratings prior to each patient’s participation in the randomized
954 	portion of the study. 	We recommend randomization occur after evaluation of the 7 total AM and
955 	PM ratings, and the randomized portion of the study can start in the morning of Day 1 after the
956 	AM baseline scoring.
957
958 	Symptom scores during the randomized treatment period would consist of the PM score on Day 1,
959 	and the 26 AM and PM ratings on Days 2 to 14, resulting in 27 total ratings. 	We recommend
960 	the study be multicenter to avoid potential investigator bias. 	A double dummy design is not
961 	recommended for study blinding of aqueous nasal sprays due to a concern that the doubled fluid
962 	volume may result in washing the drug from its nasal deposition sites, potentially resulting in an
963 	altered safety and efficacy profile. 	However, study blinding is a critical consideration, and we
964 	recommend a description of how the T, R and P products are to be masked be carefully described
965 	in the study protocol.
966
967 	We recommend the equivalence analysis be conducted as an evaluable (per protocol) analysis
968 	rather than an intent-to-treat analysis. 	The evaluable population would consist of compliant
969 	patients who missed no more than a specified number of days of symptom scores, took no
970 	contraindicated concurrent medications, and had no protocol violations. 	The protocol would
971 	describe the specific criteria used to exclude randomized subjects, resulting in the reduced subset
972 	of subjects for analysis (FDA Guideline for the Format and Content of the Clinical and
973 	Statistical Sections of an Application, Section III.B.9). In addition to the equivalence analysis,
974 	an efficacy analysis would be conducted to demonstrate study sensitivity to the T and R
975 	products. The efficacy analysis would be conducted as an intent-to-treat analysis, and the intent-
976 	to-treat population would be clearly defined. Because specific study recommendations are not
977 	provided in this guidance, we recommend a protocol for a BE study with a clinical endpoint for a
978 	specific suspension drug product be submitted prior to the conduct of the study to the appropriate
979 	review division at FDA.
980
981 	D. 	Clinical BE Study Endpoints
982
983 	The endpoints for the equivalence and efficacy analyses should be patient self-rated TNSS.
984 	These most often include a composite score of runny nose, sneezing, nasal itching, and
985 	congestion, although addition of non-nasal symptoms to the composite score maybe pertinent
986 	for certain drug products.15 TNSS is a categorical variable, classified into a number of discrete
987 	categories, as opposed to a continuous variable. 	A common allergic rhinitis rating system uses
988 	a four-point scale with signs and symptoms ordered in severity from 0 (no symptoms) to 3
989 	(severe symptoms), as follows16 :
990
991 	• 	0 = absent symptoms (no sign/symptom evident)
15 Draft guidance Allergic Rhinitis: Clinical Development Programs for Drug Products, was issued in April 2000,
once finalized it will represent the Agency’s thinking on this topic.
16 Other scoring systems were proposed in the draft guidance Allergic Rhinitis: Clinical Development Programs for
Drug Products April 2000. Once finalized, it will represent the Agency’s thinking on this topic.
23

Contains Nonbinding Recommendations
Draft — Not for Implementation
992 	• 	1 = mild symptoms (sign/symptom clearly present, but minimal awareness; easily
993 	tolerated)
994 	• 	2 = moderate symptoms (definite awareness of sign/symptom that is bothersome but
995 	tolerable)
996 	• 	3 = severe symptoms (sign/symptom that is hard to tolerate; causes interference with
997 	activities of daily living and/or sleeping)
998
999 	We recommend the endpoints for the equivalence and efficacy analyses be expressed as mean
1000 	change from baseline (pretreatment) of the TNSS, expressed in absolute units, rather than
1001 	percent change from baseline. The study report would include the daily AM and PM 12-hour
1002 	reflective symptom scores. In addition, the report would include the mean symptom score over
1003 	the 7 total AM and PM ratings of the placebo run-in period and the mean symptom score over
1004 	the 27 ratings of the randomized treatment period. 	For the equivalence and efficacy analyses,
1005 	the primary endpoint would be reflective scores for the 12-hour pooled TNSS over the two
1006 	week randomized portion of the study. 	However, instantaneous scores would also be
1007 	provided as a secondary endpoint. Statistical approaches for analysis of the rhinitis study data
1008 	are provided in Appendix F.
1009
1010 	Safety assessments would be made before (at screening or baseline) and at end-of-treatment.
1011 	Adverse events would be reported daily.
1012
1013
1014 	VII. 	PK STUDIES FOR SYSTEMIC EXPOSURE
1015
1016 	A. 	General Information
1017
1018 	The Agency recommends that plasma concentration-time profiles from BA and BE studies be
1019 	used to evaluate systemic exposure for suspension drug products that produce sufficiently high
1020 	concentrations of the moiety(ies) to be measured to allow reliable analytical measurement for an
1021 	adequate length of time after nasal administration. 	The recommended moiety(ies) to be
1022 	measured in the BA and BE studies are described elsewhere.17
1023
1024 	Systemic drug levels that occur with locally acting drug products are generally in the low ng/mL
1025 	or low pg/mL range, depending on the drug and the drug product. 	Validated bioanalytical
1026 	methodology may be available for many of the nasal corticosteroid drugs. For these drugs, pilot
1027 	studies are not needed prior to conducting the full-scale PK study. 	If validated methodology is
1028 	unavailable, a small-scale, single-dose pilot study, or when appropriate, a small-scale, multiple-
1029 	dose pilot study, may be helpful in assessing the proposed analytical methodology and
1030 	determining whether sufficiently high drug concentrations are attained. 	A PK study for
1031 	systemic exposure would be preferred to a PD or clinical study for systemic absorption (Section
1032 	VIII). If a sponsor has convincing data based on unsuccessful attempts to conduct the PK study
1033 	in order for a PD or clinical study for systemic absorption could be used. If systemic exposure
1034 	were established based on a PK study, a PD or clinical study for systemic absorption (Section
17 Guidance for Industry, Bioavailability and Bioequivalence Studies for Orally Administered Drug Products -
General Considerations (October 2000). Once finalized it will represent the Agency’s thinking on this topic.
24

Contains Nonbinding Recommendations
Draft — Not for Implementation
1035 	VIII) would not be requested.
1036
1037 	B. 	Study Batches
1038
1039 	The Agency recommends that the BA batch used for the PK systemic exposure study be a
1040 	pivotal clinical trial batch. 	Alternatively, a PK batch similar to the batch used in a pivotal
1041 	clinical trial can be used, in which case we recommend that any differences between the PK
1042 	batch and the pivotal clinical trial batch be discussed with the appropriate CDER review
1043 	division prior to the study. 	If the PK batch is not one of the three batches used for the in
1044 	vitro BA studies (Section V.A.1), make sure that in vitro BA data are provided for the PK
1045 	batch using the same protocols as for the three batches.
1046
1047 	For a BE study, the batches of T and R would be the same batches used for the clinical study for
1048 	local delivery, and each of these batches would be one of the three batches used for the in vitro
1049 	BE studies.
1050
1051 	C. 	Study Design and Subject Inclusion Criteria
1052
1053 	The BA study to characterize systemic exposure can be one of the same PK studies conducted to
1054 	address clinical pharmacology and biopharmaceutics questions of regulatory interest. 	The BA
1055 	study can be conducted in healthy subjects or allergic rhinitis (AR) patients. 	Where
1056 	appropriate, the BA study would include a reference product that may be an oral or intravenous
1057 	solution, oral suspension, or other nasal product. 	Consultation with the appropriate review
1058 	division is recommended regarding whether a comparative or noncomparative BA study is
1059 	appropriate.
1060
1061 	For an NDA or an ANDA, the in vivo BE study would be conducted with a replicate or
1062 	nonreplicate randomized crossover design. 	For aqueous nasal sprays, the study would be
1063 	conducted at the maximum labeled adult dose to maximize plasma drug levels, while avoiding
1064 	the possibility of alteration of the drug deposition pattern within the nose at higher volumes
1065 	when dosed above label claim. 	The deposition pattern could be altered due to loss of drug
1066 	from the nasal cavity at these higher volumes, due either to drainage into the nasopharynx or
1067 	externally from the nasal cavity. 	Although alteration of the deposition pattern may be less
1068 	likely for a nasal aerosol when dosed above the maximum labeled number of actuations, the
1069 	same study design and dose as for aqueous nasal sprays would be followed. 	We recommend
1070 	that subjects for the study be healthy, with exclusions primarily for reasons of safety. 	The
1071 	study protocol would include information regarding time interval between doses to each nostril
1072 	and subject head position during dosing.
1073
1074 	This guidance recommends that the PK study generally be conducted as a single-dose study.
1075 	Such studies are more sensitive than multiple dose studies in assessing rate of release of the drug
1076 	substance from the drug product into the systemic circulation. 	In addition, the nasally dosed
1077 	corticosteroids tend to have biologic half-lives ranging from less than one hour up to about eight
1078 	hours. For these products, when dosed either once or twice daily, systemic accumulation is
1079 	expected to be relatively low, thus a multiple dose study may not result in a more reliable
1080 	analytical measurement. 	However, there may be drugs that, due to pharmacokinetic
25

Contains Nonbinding Recommendations
Draft — Not for Implementation
1081 	characteristics, yield higher concentrations in a multiple-dose study, enabling the drug
1082 	moiety(ies) of interest to be measured more reliably than in a single-dose study. 	For these
1083 	drugs, a multiple-dose PK study would be preferred to a single-dose study.
1084
1085 	D. 	Study Measure s
1086
1087 	The following BA and BE measures are considered pivotal17 in a single-dose study: AUC0 - tlast (a
1088 	measure of total exposure); AUC0 - 4 (a measure of total exposure); and C max (peak exposure). 	If
1089 	AUC0 - 4 cannot be determined reliably due to inability to estimate k el accurately, total exposure
1090 	would be based only on AUC0 - tlast. 	The following BA and BE measurements and plasma
1091 	concentrations provide supportive PK characterization: plasma concentrations at each sampling
1092 	time; Tmax ; and kel . 	The following BA and BE measurements are considered-pivotal for a
1093 	multiple-dose study: AUC0 - t(total exposure), where t is the dosing interval; and C max (peak
1094 	exposure). 	Tmax data should also be provided as supportive characterization.
1095
1096 	Statistical analysis information is provided in Appendix G.
1097
1098
1099 	VIII. 	PD OR CLINICAL STUDIES FOR SYSTEMIC ABSORPTION
1100
1101 	A. 	General Information
1102
1103 	As stated in Section VI.A, at present only certain corticosteroids are formulated as suspension
1104 	products and require product quality in vivo studies. 	For those suspension drug products for
1105 	which the moiety(ies) to be measured in the blood or plasma (Section VII) are too low to allow
1106 	reliable analytical measurement for an adequate length of time, PD or clinical endpoint studies
1107 	serve as measures of systemic absorption (Section II.A.2). However, PK studies as measures of
1108 	systemic exposure are preferred if at all possible. 	As stated in Section VII, if a sponsor has
1109 	convincing data based on unsuccessful attempts to conduct the PK study a PD or clinical study
1110 	would be used in lieu of the PK study. 	The BA study to characterize systemic absorption may
1111 	be one of the same clinical studies conducted to establish the safety of the drug product. 	The
1112 	study would be conducted under an authorized IND in support of a forthcoming NDA (21 CFR
1113 	314.126).
1114
1115 	If a PD or clinical study is to be conducted (see previous paragraph), the recommended systemic
1116 	absorption BE study design for nasal corticosteroids would be assessment of the HPA axis.
1117 	The study would be conducted at the maximum labeled adult dose of the nasal aerosol or nasal
1118 	spray to maximize study sensitivity. 	However, the study design would be based on an
1119 	understanding that the maximum labeled dose over a 6-week period (Section VIII.C) may not
1120 	result in detectable adrenal suppression by T and R because this dose may be at or near the
1121 	bottom of the adrenal suppression dose-response curve. 	In addition to a test product placebo
1122 	(P), we recommend an active control such as prednisone be included to ensure that the study is
1123 	sufficiently sensitive to detect a drug effect (sensitivity analysis). 	Ensure that the active control
1124 	dose is sufficiently large and the duration sufficiently long to produce a statistically significant
1125 	response relative to placebo, with a duration sufficiently short to minimize undue exposure or
1126 	risk to subjects. 	Determination of the optimum active control dose and dosing regimen may
26

Contains Nonbinding Recommendations
Draft — Not for Implementation
1127 	call for a pilot study by the sponsor. The pilot study may determine that an initial phase of the
1128 	6-week study period may use a matching active control placebo, with active control given over
1129 	the remainder of the study period, in an effort to reduce patient exposure to the active control.
1130 	The pilot study can also provide an estimate of the number of subjects to be included in the
1131 	pivotal study to yield a statistically significant difference in the HPA axis endpoint between the
1132 	active control and the test product placebo (i.e., the aerosol or spray placebo). 	It may also
1133 	allow estimation of the number of subjects to be included to characterize any HPA axis effects
1134 	or lack thereof and to allow conclusions about any relative effects of T versus P and R versus P
1135 	(“relative assessment of the HPA axis”; Appendix G.B). Conduct of the study in allergic rhinitis
1136 	(AR) patients will allow an efficacy assessment to evaluate compliance with the study protocol
1137 	(efficacy analysis). 	Therefore, AR patients, rather than healthy, non-allergic patients are
1138 	recommended as the study population. 	We also recommend that other measures of compliance
1139 	be instituted, including before and after weighing of the aerosol or spray container and diary
1140 	entry of drug use.
1141
1142 	Because this section does not provide specific recommendations, we recommend sponsors
1143 	submit prior to the conduct of the study a protocol for a BE study with a PD or clinical endpoint
1144 	for a specific drug product to the appropriate review division at FDA. For an NDA, the same
1145 	adequate and well-controlled clinical trials in humans conducted under an authorized IND, used
1146 	to establish the safety and effectiveness of a drug product in support of a forthcoming NDA (21
1147 	CFR 314.126), can be used in some cases to establish BA or, when comparative, BE (21 CFR
1148 	320.24). For an ANDA, if the maximum single or total daily dose of the active control in the
1149 	pilot or full-scale study exceeds that specified in the labeling of the selected active control drug
1150 	product, an authorized Bio-IND will be needed.13
1151
1152 	B. 	Clinical Study Batches
1153
1154 	The Agency recommends the BA batch used for the study be a pivotal clinical trial batch used in
1155 	the in vitro BA studies (Section V.A). For BE studies for an NDA, the batches of T and R would
1156 	be batches used in in vitro testing. 	For an ANDA, the batches of T and R used for the systemic
1157 	absorption study would be the same batches used for the clinical study for local delivery. 	Each
1158 	of these batches would be one of the three batches used for the in vitro BE studies. 	Formulation
1159 	and device recommendations for the P are described in Section VI.B. An active control such as
1160 	prednisone is recommended. 	For blinding, matching active control placebo (identical in
1161 	appearance to the active control) is also recommended.
1162
1163 	C. 	Clinical BE Study Designs and Subject Inclusion Criteria
1164
1165 	We recommend the study be conducted as a placebo and active-controlled, randomized, double-
1166 	blind, parallel design comparing T and R for a 6-week duration. The study would not be
1167 	conducted as a subset of the 2-week local delivery rhinitis study (Section VI). Subjects would be
1168 	patients with a history of AR. The relative assessment of HPA axis suppression would be
1169 	conducted as an evaluable (per protocol) analysis. 	The sensitivity analysis and efficacy analysis
1170 	would be conducted as intent-to-treat analyses. 	The protocol would specify whether placebo
1171 	responders will or will not be excluded from the analysis. 	We recommend that subjects be
1172 	domiciled within the clinical study center during the days of HPA axis assessment. 	Domiciling
27

Contains Nonbinding Recommendations
Draft — Not for Implementation
1173 	the subjects during the 24-hour urine or plasma collection periods can help to conduct the study-
1174 	related procedures reliably and completely. 	T and R would be dosed at the maximum labeled
1175 	adult dose. P would be dosed at the same frequency and number of actuations per nostril as T
1176 	and R. 	As stated above, the study would include an active control such as prednisone. 	Four
1177 	study arms would be included: T, R, P, and the active control. 	The randomized portion of the
1178 	study would be conducted according to a double-blinding design (i.e., all subjects would receive
1179 	both the active control (either the active control itself or a matching placebo of the active control)
1180 	and a spray or aerosol (either active or placebo)). 	The four treatment groups would be T plus
1181 	matching active control placebo, R plus matching active control placebo, P plus matching active
1182 	control placebo, and P plus active control. 	The matching active control placebo would be dosed
1183 	on days when the active control is not taken, including the placebo run-in period. 	We
1184 	recommend the number of centers conducting the HPA assessment be kept to a minimum to avoid
1185 	center-to-center variability. 	A double-dummy design is not recommended for aqueous nasal
1186 	sprays, as explained in Section VI.C. However, study blinding is a critical consideration, and we
1187 	recommend a description of how the T, R and P products are to be masked be carefully described
1188 	in the study protocol.18
1189
1190 	The expected effect for the active control would be far larger than that for the T and R products.
1191 	The sample size of the active control arm group may therefore be smaller in size than for the
1192 	other study arms. 	We recommend the sample size for the T and R study arms be sufficient to
1193 	characterize any HPA axis effects or lack thereof to allow conclusions about any relative effects
1194 	of T versus P and R versus P, as stated in Section VIII.A.
1195
1196 	We recommend timed urine or plasma samples for determination of 24-hour urinary free cortisol
1197 	(UFC) or 24-hour plasma cortisol levels, respectively, be collected. 	Collections would be made
1198 	prior to dosing (baseline) and during the last 24 hours of the 42 days of dosing (i.e., over the day
1199 	41 – 42 period) while the drug is being actively dosed.
1200
1201 	D. 	Clinical BE Study Endpoints for Corticosteroids
1202
1203 	Whether the drug is labeled for once or twice daily dosing, the endpoint can be either 24-hour
1204 	urinary free cortisol (UFC), based on a full 24-hour urine collection, or plasma cortisol levels
1205 	collected every 4 hours over a 24-hour period, with exclusion of the middle of the night
1206 	sample. For the UFC endpoint, urinary creatinine would also be measured to confirm
1207 	completeness of the 24-hour collection. The UFC value would not be corrected for
1208 	creatinine. 	We recommend for the plasma cortisol endpoint, both AUC(0-24) and the trough
1209 	(maximum effect) concentration during the dosing interval should be determined. 	The
1210 	sensitivity analysis endpoint would be baseline-adjusted prior to analysis. 	Raw data would
1211 	be provided for the relative assessment of HPA axis suppression. Efficacy analysis TNSS
1212 	data would be expressed as change from baseline.
1213
1214 	Statistical approaches for each of the analyses are provided in Appendix G.B.
18 A draft guidance entitled Allergic Rhinitis: Clinical Development Programs for Drug Products was issued in
April 2000. Once finalized, this guidance will represent the agency's thinking on this topic.
28

Contains Nonbinding Recommendations
Draft — Not for Implementation
1215
1216
1217 	IX. 	NUMBER OF RESERVE SAMPLES FOR BA AND BE TESTING
1218
1219 	Reserve samples must be retained for BA and BE studies (21 CFR 320.38 and 320.63) conducted
1220 	in vivo or in vitro. The regulations state that each reserve sample must consist of a sufficient
1221 	quantity of samples to permit FDA to perform five times all of the release tests required in the
1222 	application or supplemental application. Dose content uniformity or spray content uniformity
1223 	release tests alone usually require 30 units (canisters or bottles) per batch. Performance of other
1224 	release tests requires additional units. The number of reserve sample units required for three
1225 	batches of T and R could exceed 1000 units (up to 250 units for each batch of T and R) based on
1226 	the five-times-quantity requirement.
1227
1228 	The Agency has determined that in lieu of the five-times-quantity requirement, the quantity of
1229 	inhalant (nasal aerosol or nasal spray) test article (T) and reference standard (R) retained for
1230 	testing and analyses be at least 50 units for each batch.19 For NDAs, three batches are needed for
1231 	BA studies. Thus, we recommend at least 50 units from each of the three batches of nasal spray
1232 	or nasal aerosol be retained. However, where the reference product is another nasal aerosol or
1233 	nasal spray, at least 50 units of that batch would also be retained. For ANDAs, at least 50 units
1234 	of each of three batches would be retained for each of T and R used in in vivo or in vitro BE
1235 	studies. For NDAs and ANDAs, if the in vivo or in vitro studies include placebo aerosols or
1236 	sprays, at least 50 units of each placebo batch would also be retained. These recommendations
1237 	apply only to nasal aerosols and nasal sprays for local action covered in this guidance and which
1238 	are marketed as multiple dose products, typically labeled to deliver 30 or more actuations per
1239 	canister or bottle. The number of reserves for nasal aerosols and nasal sprays delivering less
1240 	than 30 actuations per canister or bottle is not addressed in this guidance. Additional information
1241 	regarding retention of BA and BE testing samples is pending.20
1242
1243
1244 	X. 	MULTIPLE STRENGTHS
1245
1246 	A small number of nasal sprays for local action are available in two strengths. 	Current
1247 	examples are (1) ipratropium bromide nasal spray, a solution formulation, and (2)
1248 	beclomethasone dipropionate nasal spray, a suspension formulation. 	Lower strengths of a
1249 	product ordinarily would achieve the lower dose per actuation using a lower concentration
1250 	formulation, without changing the actuator and metering valve or pump (other than diptube due
1251 	to different volumes of product or other factors) used in the higher strength product. 	The
1252 	following sections describe recommended BA and BE studies for low strengths of nasal sprays
1253 	for which BA or BE for the higher strengths has previously been established. Recommendations
1254 	are also provided for cases in which BA or BE is initially established on the low-strength
1255 	product. 	No approved nasal aerosols are available in multiple strengths, thus BA and BE
19 Quantity of Reserve Samples, Preamble to final rule, Retention of Bioavailability and Bioequivalence Testing
Samples, 58 FR 25918-26, 1993, IIC21.
20 A draft guidance for industry entitled Handling and Retention of BA and BE Testing Samples was issued in August
2002. Once finalized, it will represent the Agency’s thinking on this topic.
29

Contains Nonbinding Recommendations
Draft — Not for Implementation
1256 	recommendations are not considered for these products.
1257
1258 	A. 	Solution Formulation Nasal Sprays
1259
1260 	We recommend the BA of lower or higher strength solution formulation nasal sprays be based on
1261 	conduct of all applicable in vitro tests described in Section V. These studies are generally
1262 	noncomparative in character. Documentation of BE between T and R products would follow the
1263 	recommendations described in Section III regarding formulation and container and closure
1264 	system. Abbreviated in vitro testing, as follows, is recommended to document BE of the low-
1265 	strength T product to the low-strength R product, provided BE of the high-strength product has
1266 	been documented.
1267
1268 	In vitro test 	High Strength 	Low Strength
1269
1270 	Single Actuation Content
1271 	Through Container Life 	B, Ea B, E
1272 	Priming and Repriming 	Yes 	Yes
1273 	Droplet Size Distribution
1274 	by Laser Diffraction 	B, E 	B
1275 	Drug in Small Particles/Droplets
1276 	by Cascade Impactor 	B 	No
o
1277 	Spray Pattern 	B 	B
1278 	Plume Geometry 	B 	N
1279 	a Beginning (B), Middle (M), End (E)
1280
1281 	With the exception of the reduced testing, the Agency recommends the same protocols and
1282 	acceptance criteria used to establish BE of the high-strength products be used for the low
1283 	strength products. 	In vivo studies are not needed for documentation of BA or BE of solution
1284 	formulation nasal sprays. Initial documentation of BE of the low-strength product would be
1285 	based on all applicable in vitro tests described in Section V. For subsequent documentation of
1286 	BE for the high-strength product, all applicable in vitro tests described above for the high-
1287 	strength product would be conducted.
1288
1289 	B. 	Suspension Formulation Nasal Sprays
1290
1291 	We recommend BA of lower strength suspension formulation nasal sprays be based on conduct
1292 	of all applicable in vitro tests described in Section V and systemic exposure studies, assuming
1293 	availability of bioanalytical methodology to allow measurement of systemic concentrations. 	In
1294 	the absence of this methodology, we suggest BA for systemic absorption be documented through
1295 	pharmacodynamic or clinical studies.
1296
1297 	BE conditions for the lower strength product would include:
1298
1299 	1. 	Documentation of BE for the high-strength test and reference products, based on
1300 	acceptable comparative formulations and container and closure systems,
1301 	comparative in vitro data, and comparative in vivo data.
30

Contains Nonbinding Recommendations
Draft — Not for Implementation
1302
1303 	2. 	Acceptable comparative formulations and container and closure systems for the
1304 	low-strength test and reference products.
1305
1306 	3. 	Acceptable comparative studies for low-strength test and reference products for
1307 	all applicable in vitro tests in Section V.
1308
1309 	4. 	Proportionally similar Single Actuation Content Through Container Life between
1310 	high- and low-dose test product and high- and low-dose reference product.
1311
1312 	In vivo studies would not be needed for documentation of BE of the lower strength products.
1313 	For cases in which an ANDA applicant initially documents BE on the low-strength suspension
1314 	formulation product, and subsequently submits an ANDA for the high-strength product, full in
1315 	vitro and in vivo documentation of BE would be provided for the high-strength product.
1316
1317
1318 	XI. 	SMALLER CONTAINER SIZE S
1319
1320 	Nasal aerosols and nasal sprays may be available in two container sizes. 	Current examples are:
1321 	(1) beclomethasone dipropionate nasal aerosol, a suspension formulation; (2) fluticasone
1322 	propionate nasal spray, a suspension formulation; and (3) cromolyn sodium nasal spray, a
1323 	solution formulation. Smaller container sizes of nasal aerosols would be formulated with the
1324 	same components and composition, metering valve, and actuator as the large container size that
1325 	was studied in pivotal clinical trials (NDA) or for which BE has been documented (ANDA).
1326 	Smaller container sizes of nasal sprays would be formulated with the same components and
1327 	composition, pump, and actuator as the large container size that was studied in pivotal clinical
1328 	trials (NDA) or for which BE has been documented (ANDA). Where this is the case, no further
1329 	documentation of either BA or BE is necessary. 	However, re-establishing proper priming,
1330 	given a change in the volume of components of the device that will be filled to deliver an
1331 	actuation, may in some cases be appropriate (Section V.B.7).
1332
1333
1334 	REFERENCES
1335
1336 	Advisory Committee for Pharmaceutical Science Meeting, “Report from the Orally Inhaled and
1337 	Nasal Drug Products Subcommittee,” Rockville, MD, Transcript, July 19, 2001, pp. 24-
1338 	91.
1339
1340 	Borgstrom L, Asking L, Beckman O, Bondesson E, K äll é n A, Olsson B. Dose Variation , within
1341 	and between individuals, with different inhalation systems. Respiratory Drug Delivery V,
1342 	Interpharm Press, Buffalo Grove, IL, 1996, pp. 19-24.
1343
1344 	Daley-Yates PT, Price AC, Sisson JR, Pereira A, Dallow N. Beclomethasone dipropionate:
1345 	absolute bioavailability, pharmacokinetics and metabolism following intravenous, oral,
1346 	intranasal and inhaled administration in man. Br J Clin Pharmacol 2001;51:400-9.
1347
1348 	Kublic H., Vidgren MT. Nasal Delivery Systems and Their Effect on Deposition and Absorption.
31

Contains Nonbinding Recommendations
Draft — Not for Implementation
1349 	Advanced Drug Delivery Reviews 1998;29:157-77.
1350
1351 	Meltzer EO, Jalowayski AA, Orgel HA, Harris AG. Subjective and objective assessments in
1352 	patients with seasonal allergic rhinitis: Effects of therapy with mometasone furoate nasal
1353 	spray. J Allergy Clin Immunol 1998;102:39-49.
1354
1355 	Newman S, Steed K, Hooper G, Käll é n A, Borgstrom L. Comparison of gamma scintigraphy and
1356 	a pharmacokinetic technique for assessing pulmonary deposition of terbutaline sulphate
1357 	delivered by pressurized metered dose inhaler. Pharm Res 1995;12:231-6.
1358
1359 	Suman JD, Laube BL, Lin T, Brouet G, Dalby R. Validity of in vitro tests on aqueous spray
1360 	pumps as surrogates for nasal deposition. Pharm Res 2002;19:1-6.
1361
1362 	Task Group on Lung Dynamics: Deposition and Retention Models for Internal Dosimetry of the
1363 	Human Respiratory Tract. Health Phys 1966;12:173-207.
1364
1365 	Welch MJ, Bronsky EA, Grossman J, Shapiro GG, Tinkelman DG, Garcia JD, Gillen MS.
1366 	Clinical evaluation of triamcinolone acetonide nasal aerosol in children with perennial
1367 	allergic rhinitis. Annals Allergy 1991;67:493-8.
1368
1369 	Williams RL, Adams W, Chen M-L, Hare D, Hussain A, Lesko L, Patnaik R, Shah V, FDA
1370 	Biopharmaceutics Coordinating Committee. Where are we now and where do we go next
1371 	in terms of the scientific basis for regulation on bioavailability and bioequivalence?
1372 	Europ J Drug Metab Pharmacokinet 2000;25:7-12.
32

Contains Nonbinding Recommendations
Draft — Not for Implementation
373
374 	TABLE 1
375 	RECOMMENDED IN VITRO STUDIES FOR BA AND BE OF NASAL AEROSOLS AND NASAL
376 	SPRAYS
377
378
TEST1 	BA AND BE
STUDY MEASURE(S)
BE MEASURE(S) FOR
STATISTICAL
EVALUATION
LIFESTAGE(S)
B (beginning), M (middle),
E (end)
STATISTICAL EVALUATION
FOR BE
PBE (population bioequivalence)
GUIDANCE
SECTIONS
Single Actuation Content
Through Container Life
Drug mass per single
actuation
Same as previous column 	B, M, E (aerosols)
B, E (sprays)
PBE 	V.B.1, App. B, C
Droplet Size Distribution
by Laser Diffraction
D10, D50, D90, span
at 2 distances
D50, span 	B, E 	PBE 	V.B.2, App. B, C
Drug in Small
Particles/Droplets by
Cascade Impactor
Drug mass below upper
stage
Same as previous column 	B (sprays) 	PBE modified to be one-sided with
respect to the mean comparison
V.B.3, App. B, D
Particle/Droplet Size
Distribution by Cascade
Impactor
Drug mass on individual
accessories, stages, etc –
profile analysis
Deposition profile 	B (aerosols) 	Profile analysis 	V.B.3, App. B, E
Drug Particle Size
Distribution by Microscopy
for Suspensions
Drug CMD;
extent of agglomerates
Same as previous column 	B 	Not applicable 	V.B.4
Spray Pattern 	Automated analysis: area,
ovality ratio at 2 distances
or
Manual analysis: Dmax ,
ovality ratio at 2 distances
Qualitative – shape
comparison
Quantitative - Same as
previous column
B 	PBE for area and ovality ratio
(automated analysis)
or
Dmax and ovality ratio manual
analysis
V.B.5, App. C
Plume Geometry 	Height, width, and cone
angle of one side view at
one delay time
Width and cone angle of
one side view at one delay
time
B 	Point estimates 	V.B.6
Priming and Repriming 	Drug mass per single
actuation at first primed or
reprimed actuation
Same as previous column
for Priming, and Repriming
if in precursor product (R)
labeling
B (Priming)
Lifestage not specified
(Repriming)
Point estimate relative to label
claim if in precursor product (R)
labeling
V.B.7
379 	1 Although alternate test methods may be appropriate for certain tests, if validated, we recommend sponsors planning to use such methods contact the appropriate reviewing division prior to use.
33