CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`APPLICATION NUMBER:
`
`207154Orig1s000
`
`STATISTICAL REVIEW(S)
`
`
`
`
`
`
`
`

`

`U.S. Department of Health and Human Services
`Food and Drug Administration
`Center for Drug Evaluation and Research
`Office of Translational Sciences
`Office of Biostatistics
`
`S T A T I S T I C A L R E V I E W A N D E V A L U A T I O N
`CLINICAL STUDIES
`
`NDA/BLA #:
`Drug Name:
`Indication(s):
`Applicant:
`Date(s):
`
`Review Priority:
`
`NDA 207154
`ACZONE (dapsone) gel 7.5%
`Acne Vulgaris
`Allergan, Inc.
`Letter Date: 4/28/2015
`PDUFA Date: 2/28/2016
`Standard
`
`Biometrics Division:
`Division of Biometrics III
`Statistical Reviewer:
`Matthew Guerra, Ph.D.
`Concurring Reviewers: Mohamed Alosh, Ph.D.
`
`Medical Division:
`Clinical Team:
`Project Manager:
`
`Division of Dermatology and Dental Products
`Patricia Brown, M.D. / Gordana Diglisic, M.D.
`Strother Dixon / Cristina Attinello
`
`Keywords: Acne vulgaris, superiority trial, multiple imputation
`
`Reference ID: 3873729
`
`

`

`Table of Contents
`EXECUTIVE SUMMARY ..................................................................................................................................3
`
`1
`
`2
`
`3
`
`INTRODUCTION ................................................................................................................................................4
`2.1
`OVERVIEW.......................................................................................................................................................4
`2.1.1
`Regulatory History..................................................................................................................................4
`2.1.2
`Clinical Studies Overview.......................................................................................................................5
`2.2
`DATA SOURCES ...............................................................................................................................................5
`STATISTICAL EVALUATION .........................................................................................................................6
`3.1
`DATA AND ANALYSIS QUALITY.......................................................................................................................6
`3.2
`EVALUATION OF EFFICACY..............................................................................................................................6
`3.2.1
`Study Design and Endpoints ...................................................................................................................6
`3.2.2
`Statistical Methodologies........................................................................................................................7
`3.2.3
`Patient Disposition, Demographics and Baseline Characteristics.........................................................8
`3.2.4
`Primary Efficacy Results.......................................................................................................................10
`3.2.5
`Secondary Efficacy Results...................................................................................................................12
`3.2.6
`Global Assessment of Acne Score (GAAS) vs. Lesion Counts at Week 12 ...........................................13
`3.3
`EVALUATION OF SAFETY...............................................................................................................................16
`3.3.1
`Extent of Exposure ................................................................................................................................16
`3.3.2
`Adverse Events......................................................................................................................................16
`FINDINGS IN SPECIAL/SUBGROUP POPULATIONS..............................................................................20
`4.1
`GENDER, RACE, AGE, AND COUNTRY............................................................................................................20
`4.2
`CENTER..........................................................................................................................................................21
`SUMMARY AND CONCLUSIONS .................................................................................................................25
`5.1
`STATISTICAL ISSUES AND COLLECTIVE EVIDENCE........................................................................................25
`5.2
`CONCLUSIONS AND RECOMMENDATIONS......................................................................................................27
`APPENDIX..................................................................................................................................................................28
`A.1
`ACNE SYMPTOM AND IMPACT SCALE (ASIS) [ITEMS 1 TO 10]..................................................................28
`SIGNATURES/DISTRIBUTION LIST....................................................................................................................30
`
`4
`
`5
`
`
`
`Reference ID: 3873729
`
`2
`
`

`

`1 EXECUTIVE SUMMARY
`
`The applicant has developed ACZONE® (dapsone) gel, 7.5% for the topical treatment of acne
`vulgaris in patients 12 years of age and older. ACZONE® (dapsone) gel, 5% was approved on
`July 7, 2005 for the indication of topical treatment of acne vulgaris. It should be noted that the
`approved dose regimen for ACZONE® (dapsone) gel, 5% is twice daily and the proposed dose
`regimen for ACZONE® (dapsone) gel, 7.5% is once daily.
`
`The applicant submitted data from two identically-designed, randomized, multicenter, vehicle-
`controlled, parallel-group, Phase 3 trials (Trials 006 and 007). For enrollment, the protocol
`specified the following key inclusion criteria: 12 years of age or older, a Global Acne
`Assessment Score (GAAS) of 3 (moderate), 20-50 inflammatory lesions (papules and pustules)
`on the face, and 30-100 non-inflammatory lesions (open comedones and closed comedones) on
`the face. The protocol-specified co-primary efficacy endpoints were the proportion of subjects
`achieving a GAAS score of 0 (none) or 1 (minimal) at Week 12 and the absolute change in
`inflammatory and non-inflammatory lesion counts from baseline to Week 12. Secondary efficacy
`endpoints included percent change in inflammatory and non-inflammatory lesion counts from
`baseline to Week 12.
`
`Table 1 presents the results of the co-primary efficacy endpoints and the secondary efficacy
`endpoints of percent change in inflammatory and inflammatory lesion counts from baseline to
`Week 12. In both trials, ACZONE gel, 7.5% was statistically superior (p-values ≤ 0.004) to
`vehicle gel for all endpoints presented in Table 1.
`
`Table 1: Results for the Co-Primary and Secondary Efficacy Endpoints at Week 12
`Trial 006
`Trial 007
`
`ACZONE
`Vehicle
`ACZONE
`Vehicle
`(N=1044)
`(N=1058)
`(N=1118)
`(N=1120)
`Endpoints
`Co-Primary:
`
`
`
`
`30%
`21%
`30%
`21%
`GAAS (none or minimal): n (%)
`
`
`
`Absolute Change in:
`16.1
`14.3
`14.0
` Inflammatory Lesions: Mean
`20.7
`18.0
`18.7
` Non-Inflammatory Lesions: Mean
`Secondary:
`
`
`
`
`
`
`Percent Change in:
`56%
`49%
`48%
` Inflammatory Lesions: Mean
`45%
`39%
`41%
` Non-Inflammatory Lesions: Mean
`Source: Reviewer’s Analysis (same as Applicant’s Analysis)
`
`54%
`46%
`
`15.6
`20.8
`
`For the assessment of GAAS, the interpretation of a “few” or “no” lesions seemed to vary from
`investigator to investigator. Some subjects counted as successes under the GAAS seemed to have
`relatively high lesion counts for the definition of “none” (no evidence of facial acne vulgaris) or
`“minimal” (a few non-inflammatory lesions (comedones) are present; a few inflammatory
`lesions (papules/pustules) may be present). Subjects scored as 0 (none) had as many as 10
`inflammatory lesions or 45 non-inflammatory lesions. Subjects scored as 1 (minimal) had as
`many as 57 inflammatory lesions or 102 non-inflammatory lesions.
`
`3
`
`Reference ID: 3873729
`
`

`

`2 INTRODUCTION
`
`2.1 Overview
`
`The applicant, Allergan, is developing ACZONE® (dapsone) gel, 7.5% for the topical treatment
`of acne vulgaris in patients 12 years of age and older. ACZONE® (dapsone) gel, 5% was
`approved on July 7, 2005 for the indication of topical treatment of acne vulgaris. It should be
`noted that the approved dose regimen for ACZONE® (dapsone) gel, 5% is twice daily and the
`proposed dose regimen for ACZONE® (dapsone) gel, 7.5% is once daily.
`
`2.1.1 Regulatory History
`
`On August 28, 2013, the Agency and the applicant met for an End-of-Phase 2 (EOP2) meeting to
`discuss the development plan for ACZONE (dapsone) gel, 7.5%. The applicant proposed to
`conduct two identically-designed Phase 3 trials (Trials 006 and 007) and submitted the protocol
`for these trials in the meeting package. The applicant proposed the co-primary efficacy endpoints
`of proportion of subjects with success on the GAAS (i.e., score of 0 or 1) at Week 12 and
`absolute change in lesion counts (inflammatory, non-inflammatory, and total) from baseline to
`Week 12. The Agency recommended that the co-primary endpoints regarding lesion counts be
`absolute change in inflammatory and non-inflammatory lesion counts from baseline to Week 12
`(i.e., not include total as a co-primary endpoint). The Agency also commented that several of the
`secondary endpoints are closely related and some of the secondary endpoints might not be
`clinically relevant for labeling. The Agency stated that the secondary endpoints of percent
`change in inflammatory and non-inflammatory lesion count from baseline to Week 12 are
`acceptable. In addition, the Agency stated that the proposed patient reported outcomes may have
`limited utility for eventual product labeling. The Agency also provided comments regarding the
`handling of missing data (i.e., recommended a more scientifically sound approach, such as
`multiple imputation or modeling approach, instead of the last observation carried forward
`(LOCF) approach).
`
`On October 7, 2013, the applicant submitted amended protocols for the Phase 3 trials proposed
`during the EOP2 meeting. An advice letter was sent to the applicant on January 15, 2014. The
`Agency reiterated the comments from the EOP2 meeting regarding the absolute change in total
`lesion counts as a co-primary endpoint and the limited utility of the proposed patient reported
`outcomes (i.e., the Acne Symptom and Impact Scale (ASIS)).
`
`On February 11, 2014, the applicant submitted their responses to the Agency’s comments
`conveyed in the advice letter sent on January 15, 2014. In addition, on February 18, 2014, the
`applicant submitted their Patient Reported Outcomes (PRO) Questions Document, a new Acne
`Symptom and Impact Scale (ASIS) PRO Dossier and a draft statistical analysis plan (SAP) for
`their pivotal Phase 3 trials. An advice letter regarding these two submissions was sent to the
`applicant on June 13, 2014. The Agency provided extensive comments regarding the ASIS. For
`any PRO endpoints that are proposed to support labeling claims, the Agency recommended pre-
`specifying an appropriate responder definition, making appropriate adjustments for multiple
`endpoints, and discussing these considerations with the Agency.
`
`4
`
`Reference ID: 3873729
`
`

`

`On November 19, 2014, the applicant and the Agency met for a Pre-NDA meeting. The Agency
`provided general comments on how the data should be submitted (data tabulation datasets, data
`definition files, annotated case report forms, and analysis datasets). The applicant notified the
`Agency that a clinical center (16078; Dr. Ellen Marmur) did not follow Good Clinical Practice
`(GCP) procedures. The applicant noted the following instances of non-compliance:
` Numerous inconsistencies in documentation indicating that Dr. Marmur conducted
`patient assessments when it was confirmed she was not present in the office
` Consenting, screening, and enrolling patients into the study, as well as efficacy and safety
`assessments, conducted by a study coordinator who was not eligible to conduct the
`assessments, per protocol, and not listed on the Investigator’s Form FDA 1572
` Lack of documentation for numerous patients who were randomized but who do not
`appear to have returned for any follow-up visits
`Due to the above issues, the applicant terminated the center from the trial and all ongoing
`subjects at the center were discontinued. The Agency stated that given the potential seriousness
`of the violations described, data from this center should not be included in the primary efficacy
`analysis. In addition, the Agency commented that the statistical analysis should follow the
`randomization; therefore, as the randomization was stratified by gender and center, the Agency
`recommended the applicant conduct the analyses stratified by both factors with and without
`pooling.
`
`
`2.1.2 Clinical Studies Overview
`
`The applicant submitted data from a two pivotal Phase 3 trials (Trials 006 and 007). An overview
`of the trials is presented in Table 2.
`
`Table 2: Clinical Study Overview
`
`Trial
`
` Treatment Arms
`Study Population
`Location
`ACZONE Gel, 7.5%
`U.S. (96 centers) &
`Aged 12 years and older,
`Canada (9 centers)
`GAAS of 3 (moderate), 20
`Vehicle Gel
`to 50 inflammatory lesions,
`ACZONE Gel, 7.5%
`U.S. (93 centers) &
`and 30 to 100 non-
`Canada (10 centers)
`inflammatory lesions
`Vehicle Gel
`*Excluding subjects from center 16078 (25 on ACZONE gel, 7.5% and 26 on vehicle gel).
`
`006
`
`007
`
`Number of
`Subjects
`1044*
`1058*
`1118
`1120
`
`Dates
`11/27/2013 –
`10/28/2014
`11/27/2013 –
`10/21/2014
`
`2.2 Data Sources
`
`This reviewer evaluated the applicant’s clinical study reports, datasets, clinical summaries, and
`proposed labeling. This submission was submitted in eCTD format and entirely electronic. The
`datasets in this review are archived at the following locations:
`\\cdsesub1\evsprod\NDA207154\0000\m5\datasets\
`
`5
`
`Reference ID: 3873729
`
`

`

`3 STATISTICAL EVALUATION
`
`3.1 Data and Analysis Quality
`
`The databases for the studies required minimal data management prior to performing analyses
`and no request for additional datasets were made to the applicant.
`
`3.2 Evaluation of Efficacy
`
`3.2.1 Study Design and Endpoints
`
`The applicant conducted two identically-designed Phase 3 trials (Trials 006 and 007). Both were
`randomized, double-blind, parallel-group, vehicle-controlled, 12-week trials investigating the
`safety and efficacy of ACZONE® (dapsone) gel, 7.5% compared to vehicle gel for the treatment
`of acne vulgaris. For enrollment, the protocol specified the following key inclusion criteria:
` Male or female 12 years of age or older
` Global Acne Assessment Score (GAAS) of 3 (moderate), see Table 3 for details on the
`GAAS
`20-50 inflammatory lesions (papules and pustules) on the face
`30-100 non-inflammatory lesions (open comedones and closed comedones) on the face
`
`
`
`
`Each trial was designed to enroll and randomize approximately 2180 subjects in a 1:1 ratio to
`either ACZONE® gel, 7.5% or vehicle gel. Randomization was stratified by gender and center.
`Subjects applied study product once daily for 12 weeks. Subjects were evaluated at the following
`study visits: screening, baseline (Day 1) and Weeks 1, 2, 4, 8, and 12.
`
`Table 3: Global Acne Assessment Score (GAAS)
`Grade
`Description
`None
`No evidence of facial acne vulgaris
`Few non-inflammatory lesions (comedones) are present; a few inflammatory
`Minimal
`lesions (papules/pustules) may be present
`Several to many non-inflammatory lesions (comedones) are present; a few
`Mild
`inflammatory lesions (papules/pustules) are present
`Moderate Many non-inflammatory (comedones) and inflammatory lesions
`(papules/pustules) are present; no nodulo-cystic lesions are allowed
`Significant degree of inflammatory disease; papules/pustules are a predominant
`feature; a few nodulo-cystic lesions may be present; comedones may be present
`
`0
`1
`
`2
`
`3
`
`4
`
`Severe
`
`The protocol specified the following co-primary efficacy endpoints:
` Proportion of subjects with a 0 (none) or 1 (minimal) on the GAAS at Week 12
` Absolute change in inflammatory lesion counts from baseline to Week 12
` Absolute change in non-inflammatory lesion counts from baseline to Week 12
`
`6
`
`Reference ID: 3873729
`
`

`

`The protocol specified the following secondary efficacy endpoints:
` Absolute change in total lesion counts from baseline to Week 12
` Percent change in total lesion counts from baseline to Week 12
` Percent change in inflammatory lesion counts from baseline to Week 12
` Percent change in non-inflammatory lesion counts from baseline to Week 12
` Proportion of subjects who report "very good" or "excellent" in Item 10 from the Acne
`Symptom and Impact Scale (ASIS) at Week 12, see Appendix for details on the ASIS
` Absolute change in ASIS Sign Domain Score (i.e., the average score over Items 1
`through 9) from baseline to Week 12
` Proportion of subjects with at least a 1-grade improvement on Item 1 from the ASIS
`(subject’s assessment of oiliness on the face) at Week 12
` Proportion of subjects with at least a 1-grade improvement on Item 8 from the ASIS
`(subject’s assessment of redness on the face) at Week 12
`However, as stated before, the Agency informed the applicant that some of secondary efficacy
`endpoints may not be relevant for labeling.
`
`3.2.2 Statistical Methodologies
`
`The primary analysis population specified in the protocol was the intent-to-treat (ITT)
`population, defined as all randomized subjects. The protocol also specified supportive analyses
`using the per-protocol (PP). The PP population was defined as all randomized subjects with no
`protocol deviations during the trial that might potentially affect the primary efficacy analyses.
`
`The statistical analysis plan (SAP) specified a pooling algorithm for centers that enrolled less
`than 24 subjects. The pooling was conducted within 5 regional areas (i.e., Canada, northeastern
`states of U.S., southern states of U.S., west coast states of US, and all other states of the U.S.).
`Within each regional area, centers were ranked in descending order based on the total number of
`subjects enrolled. The first center with fewer than 24 subjects is combined with the next center,
`or with more centers if needed, until the total number in the pooled center reaches or exceeds 24
`subjects. The algorithm continues down the list, and if the last pooled center has less than 24
`subjects, then the last pooled center is combined with the previous pooled center.
`
`The protocol-specified analysis method for the co-primary efficacy endpoint of GAAS success
`(i.e., none or minimal) at Week 12 was the Cochran-Mantel-Haenszel (CMH) test stratified by
`gender. The protocol specified investigating the treatment-by-gender interaction using the
`Breslow-Day test at α = 0.10 level. The SAP (finalized after the protocol) specified investigating
`the treatment-by-center interaction using the Breslow-Day test at α = 0.10 level.
`
`For the co-primary efficacy endpoints of absolute change in inflammatory and non-inflammatory
`lesion counts from baseline to Week 12, the protocol-specified analysis method was analysis of
`covariance (ANCOVA) with treatment, baseline lesion counts, and gender in the model. As a
`sensitivity analysis, the protocol specified including the treatment-by-center (pooled) interaction.
`If significant at the 0.10 level, the protocol specified that “data will be further explored by
`excluding those investigational centers with a large number of deviations.”
`
`7
`
`Reference ID: 3873729
`
`

`

`For the analysis of the secondary efficacy endpoints of absolute change in total lesion counts and
`percent change in lesion counts (total, inflammatory, non-inflammatory) from baseline to Week
`12, the protocol specified using the same method (i.e., ANCOVA model) used to analyze the co-
`primary efficacy endpoints of absolute change in inflammatory and non-inflammatory lesion
`counts from baseline to Week 12. The protocol specified analyzmg the binary secondary efficacy
`endpoints using the CMH test stratified by gender. For the secondary endpoint of absolute
`change in ASIS Sign Domain Score from baseline to Week 12, the protocol-specified analysis
`method was ANCOVA using rank data with treatment and gender in the model.
`
`To control the Type I error rate for testing multiple secondary efficacy endpoints, the SAP
`specified using a sequential gatekeeping approach. The secondary efficacy endpoints were
`analyzed in the order specified in Section 3.2.1 of this review.
`
`For co-primary eflicacy endpoints (i.e., GAAS and lesion counts), the primary imputation
`method for the handling of missing data specified in the SAP was the multiple imputation GVJI)
`approach. Missing data was imputed using regression models with treatment, age, gender,
`baseline lesion counts (only for inflammatory and non-inflammatory endpoints) and previous
`visits results (e.g., missing data for Week 4 was imputed using the data from Weeks 2 and l).
`The SAP specified using the last observation cam'ed forward @OCF) as a sensitivity analysis for
`the handling of missing data. For the secondary endpoints based on the ASIS, the SAP specified
`imputing missing data using LOCF.
`
`3.2.3 Patient Disposition, Demographics and Baseline Characteristics
`
`Trial 006 enrolled and randomized a total (excluding center 16078) of 2102 subjects (1044 to
`ACZONE and 1058 to vehicle) from 105 centers (96 in US. and 9 in Canada). Trial 007 enrolled
`and randomized a total of 2238 subjects (1118 to ACZONE and 1120 to vehicle) from 103
`centers (93 in US. and 10 in Canada). Table 4 presents the disposition of subjects in Trials 006
`and 007. For Trial 006, the rate of discontinuation was slightly higher in the ACZONE arm
`comparted to the vehicle arm. For Trail 007, the rates of discontinuation were almost identical.
`
`Table 4: Dis I osition of Sub'ects T
`
`
`
`
`
`
`
`
`Trial 007
`Trial 006(1)
`Vehicle
`ACZONE
`ACZONE
`Vehicle
`
`
`
`(N=l044)
`(N=1058)
`(N=1118)
`(N=1120)
`
`
`96 (9.2%)
`82 (7.8%)
`92 (8.2%)
`93 (8.3%)
`Discontinued
`
`
`
`Adverse Event
`4 (0.4%)
`5 (0.5%)
`2 (0.2%)
`2 (0.2%)
`
`Lack ofEflicacy
`0
`1 (0.1%)
`l (0.1%)
`1 (0.1%)
`Lost to Follow-Up
`38 (3.6%)
`29 (2.7%)
`45 (4.0%)
`40 (3.6%)
`
`Other
`28 (2.7%)
`18 (1.7%)
`25 (2.2%)
`28 (2.5%)
`Personal Reasons
`2] (2.0%)
`20 (1.9%)
`15 (1.3%)
`19 (1.7%)
`
`Pregnancy
`3 (0.3%)
`3 (0.3%)
`2 (0.2%)
`l (0.1%)
`
`Protocol Violations
`2 (0.2%)
`6 (0.6%)
`2 (0.2%)
`2 (0.2%)
`
`Source: Reviewer’s Analysis (same results as Applicant’s Analysis)
`(1) Excluding subjects from cane: 16078 (a total of 51 subjecs).
`
`
`
`
`
`
`
`
`
`
`
`
`
`Reference ID: 3873729
`
`

`

`Table 5 presents the demographic and baseline disease characteristics. The demographics and
`baseline disease characteristics were generally balanced across the treatment arms within each
`trial and similar between each trial. For enrollment, the protocol specified subjects to have a
`GAAS of 3 (moderate), 20 to 50 inflammatory lesions, and 30 to 100 non-inflammatory lesions.
`One subject in Trial 006 had a GAAS of 4 (severe). In Trial 006, one subject had a baseline non-
`inflammatory lesion count less than 30 (i.e., 5 lesions), one subject had a baseline non-
`inflammatory lesion count greater than 100 (i.e., 106 lesions), and one subject had a baseline
`inflammatory lesion count less than 20 and a baseline non-inflammatory lesion count less than
`30 (i.e., 11 inflammatory lesions and 4 non-inflammatory lesions). In Trial 007, four subjects had
`a baseline inflammatory lesion count greater than 50 (i.e., 52, 53, 57), and 62 lesions, two
`subjects had a baseline non-inflammatory lesion count less than 30 (i.e., 21 and 28 lesions), and
`one subject had a baseline non-inflammatory lesion count greater than 100 (i.e., 112 lesions).
`
`
`
`Table 5: Demographics and Baseline Disease Characteristics (ITT)
`Trial 006(1)
`ACZONE
`Vehicle
`(N=1044)
`(N=1058)
`
`
`20.0 (7.4)
`20.0 (7.5)
`17.0
`17.0
`12 – 63
`12 – 53
`525 (50%)
`554 (52%)
`519 (50%)
`504 (48%)
`
`Age
` Mean (SD)
` Median
` Range
` 12-17
` 18+
`Gender
` Male
` Female
`Race
` White
` Black
` Hispanic
` Asian
` Other
`Country
` U.S.
` Canada
`GAAS
` 3 – Moderate
` 4 - Severe
`Inflammatory Lesion Counts
` Mean (SD)
` Median
` Range
`Non-Inflammatory Lesion Counts
`46.9 (16.6)
` Mean (SD)
`41.0
` Median
`4 – 100
` Range
`Source: Reviewer’s Analysis (same results as Applicant’s Analysis)
`SD: Standard Deviation
`(1) Excluding subjects from center 16078 (a total of 51 subjects).
`
`453 (43%)
`591 (57%)
`
`647 (62%)
`173 (17%)
`135 (13%)
`44 (4%)
`45 (4%)
`
`984 (94%)
`60 (6%)
`
`1043
`1
`
`28.8 (8.0)
`26.0
`11 – 50
`
`Reference ID: 3873729
`
`Trial 007
`ACZONE
`Vehicle
`(N=1118)
`(N=1120)
`
`
`20.5 (8.2)
`20.4 (7.4)
`18.0
`18.0
`12 – 61
`12 – 54
`541 (48%)
`530 (47%)
`577 (52%)
`590 (53%)
`
`476 (45%)
`582 (55%)
`
`623 (59%)
`189 (18%)
`156 (15%)
`43 (4%)
`47 (4%)
`
`997 (94%)
`61 (6%)
`
`1058
`0
`
`29.3 (8.1)
`27.0
`20 – 50
`
`500 (45%)
`618 (55%)
`
`601 (54%)
`230 (21%)
`212 (19%)
`37 (3%)
`38 (3%)
`
`1057 (95%)
`61 (5%)
`
`1118
`0
`
`29.6 (7.7)
`28.0
`20 – 62
`
`48.6 (17.5)
`43.0
`30 – 106
`
`46.7 (15.3)
`42.0
`21 – 112
`
`489 (44%)
`631 (56%)
`
`619 (55%)
`220 (20%)
`191 (17%)
`44 (4%)
`46 (4%)
`
`1058 (94%)
`62 (6%)
`
`1119
`0
`
`30.0 (7.9)
`28.0
`20 – 57
`
`46.7 (15.0)
`42.0
`30 – 100
`
`9
`
`

`

`3.2.4 Primary Efficacy Results
`
`ACZONE gel, 7.5% was statistically superior (p-values ≤ 0.004) to vehicle gel on all three co-
`primary efficacy endpoints in both trials. The results from the ITT and PP analyses were similar
`and are presented in Tables 6 and 7, respectively.
`
`Table 6: Results for the Co-Primary Efficacy Endpoints at Week 12 (MI, ITT)
`Trial 006(1)
`Trial 007
`
`Vehicle
`Vehicle
`(N=1058)
`(N=1120)
`
`P-value
`
`ACZONE
`(N=1118)
`
`ACZONE
`(N=1044)
`
`Endpoint
`GAAS:
`None or Minimal*
`Absolute Change in
`Inflammatory Lesion
`Counts:
` Mean*
` LS Mean(3)
`Absolute Change in
`Non-Inflammatory
`Lesion Counts:
`18.0
`20.7
` Mean*
`17.6
`20.8
` LS Mean(3)
`Source: Reviewer’s Analysis (same results as Applicant’s Analysis)
`*The values displayed are the averages over the 20 imputed datasets (MI).
`(1) Excluding subjects from center 16078 (a total of 51 subjects).
`(2) P-value from a CMH test stratified by gender.
`(3) LS means and p-values from an ANCOVA model with terms for treatment, gender, and baseline lesion counts.
`
`311.9 (30%)
`
`224.2 (21%)
`
`<0.001(2)
`
`333.3 (30%)
`
`234.1 (21%)
`
`
`16.1
`16.1
`
`
`14.3
`14.1
`
`
`
`<0.001(3)
`
`<0.001(3)
`
`
`15.6
`15.6
`
`20.8
`20.7
`
`
`14.0
`13.8
`
`18.7
`18.5
`
`P-value
`
`<0.001(2)
`
`
`
`<0.001(3)
`
`0.004(3)
`
`Table 7: Results for the Co-Primary Efficacy Endpoints at Week 12 (PP)
`Trial 006(1)
`Trial 007
`
`Vehicle
`Vehicle
`(N=887)
`(N=955)
`
`P-value
`
`ACZONE
`(N=950)
`
`ACZONE
`(N=880)
`
`Endpoint
`GAAS:
`None or Minimal
`Absolute Change in
`Inflammatory Lesion
`Counts:
` Mean
` LS Mean
`Absolute Change in
`Non-Inflammatory
`Lesion Counts:
`18.2
`21.1
` Mean
`17.9
`21.0
` LS Mean
`Source: Reviewer’s Analysis (same results as Applicant’s Analysis)
`(1) Excluding subjects from center 16078 (a total of 51 subjects).
`(2) P-value from a CMH test stratified by gender.
`(3) LS means and p-values from an ANCOVA model with terms for treatment, gender, and baseline lesion counts.
`
`272 (31%)
`
`199 (22%)
`
`<0.001(2)
`
`291 (31%)
`
`203 (21%)
`
`16.2
`16.1
`
`14.6
`14.4
`
`<0.001(3)
`
`<0.001(3)
`
`15.8
`15.8
`
`21.4
`21.1
`
`14.4
`14.2
`
`18.9
`18.7
`
`Reference ID: 3873729
`
`P-value
`
`<0.001(2)
`
`0.001
`
`<0.001(3)
`
`10
`
`

`

`Table 8 provides the number of subjects with missing data for the co-primary efficacy endpoints
`by week and treatment arm for both trials. In both trials, the proportion of subjects with missing
`data at Week 12 was slightly higher (9% vs. 8%) in the ACZONE arm compared to the vehicle
`arm.
`
`
`
`Table 8: Missing Data for the Co-Primary Efficacy Endpoints by Week (ITT)
`Trial 006
`Trial 007
`ACZONE
`Vehicle
`ACZONE
`Vehicle
`(N=1118)
`(N=1120)
`(N=1044)
`(N=1058)
`0 (0%)
`0 (0%)
`0 (0%)
`1 (0.1%)
`Baseline
`68 (7%)
`71 (7%)
`67 (6%)
`70 (6%)
`Week 1
`65 (6%)
`61 (6%)
`72 (6%)
`63 (6%)
`Week 2
`38 (4%)
`43 (4%)
`50 (4%)
`50 (4%)
`Week 4
`89 (9%)
`72 (7%)
`77 (7%)
`85 (8%)
`Week 8
`95 (9%)
`85 (8%)
`96 (9%)
`94 (8%)
`Week 12
`Source: Reviewer’s Analysis
`
`For all three co-primary efficacy endpoints, the primary imputation method was the multiple
`imputation approach using a regression model with treatment, age, gender, baseline lesion counts
`(only for inflammatory and non-inflammatory lesion counts) and previous visits results in the
`model (MI-Reg). The SAP also specified using LOCF as a sensitivity analysis for the handling of
`missing data. For the co-primary efficacy endpoint of IGA success, this reviewer conducted a
`sensitivity analysis where missing data was imputed as failures. In addition, for all three co-
`primary efficacy endpoints, this reviewer conduct an additional sensitivity analysis where
`missing data was imputed using the multiple imputation Markov Chain Monte Carlo (MI-
`MCMC) approach.
`
`Tables 9, 10, and 11 present the results for the co-primary efficacy endpoints in both trials by the
`various imputations methods. For both trials, the results were generally similar across the various
`methods for handling missing data.
`
`Table 9: Comparison of Different Approaches for Handling Missing Data for GAAS
`Success(1) at Week 12 (ITT)
`
`Imputation
`Method
`MI-Reg(3) (Primary)
`LOCF
`Failure
`MI-MCMC(3)
`Source: Reviewer’s Analysis
`(1) Success is defined as achieving a GAAS of 0 (none) or 1 (minimal).
`(2) P-value based on a CMH test stratified by gender.
`(3) The rates displayed are the averages of the 20 imputed datasets.
`
`Trial 007
`Vehicle
`(N=1120)
`234.1 (21%)
`218 (19%)
`215 (19%)
`232.2 (21%)
`
`P-value(2)
`<0.001
`<0.001
`<0.001
`<0.001
`
`ACZONE
`(N=1044)
`311.8 (30%)
`288 (28%)
`284 (27%)
`307.2 (29%)
`
`Trial 006
`Vehicle
`(N=1058)
`222.6 (21%)
`212 (20%)
`207 (20%)
`222.7 (21%)
`
`P-value(2)
`<0.001
`<0.001
`<0.001
`<0.001
`
`ACZONE
`(N=1118)
`333.3 (30%)
`312 (28%)
`306 (27%)
`330.5 (30%)
`
`Reference ID: 3873729
`
`11
`
`

`

`Table 10: Comparison of Different Approaches for Handling Missing Data for
`Inflammatory Lesion Counts at Week 12 (ITT)
`
`Trial 006
`Vehicle
`Imputation
`(N=1058)
`Method
`MI-Reg(3) (Primary)
`14.6
`LOCF
`13.7
`MI-MCMC(3)
`14.1
`Source: Reviewer’s Analysis
`(1) P-values from an ANCOVA model with terms for treatment, gender, and baseline lesion counts.
`(2) The rates displayed are the averages of the 20 imputed datasets.
`
`P-value(1)
`<0.001
`<0.001
`<0.001
`
`ACZONE
`(N=1118)
`15.8
`14.9
`15.4
`
`Trial 007
`Vehicle
`(N=1120)
`14.4
`13.4
`13.8
`
`P-value(1)
` 0.001
`<0.001
`<0.001
`
`ACZONE
`(N=1044)
`16.2
`15.5
`16.0
`
`Table 11: Comparison of Different Approaches for Handling Missing Data for Non-
`Inflammatory Lesion Counts at Week 12 (ITT)
`
`Trial 006
`Vehicle
`Imputation
`(N=1058)
`Method
`MI-Reg(2) (Primary)
`18.2
`LOCF
`17.4
`MI-MCMC(2)
`17.8
`Source: Reviewer’s Analysis
`(1) P-values from an ANCOVA model with terms for treatment, gender, and baseline lesion counts.
`(2) The rates displayed are the averages of the 20 imputed datasets.
`
`P-value(1)
`<0.001
`<0.001
`<0.001
`
`ACZONE
`(N=1118)
`21.4
`19.9
`20.5
`
`Trial 007
`Vehicle
`(N=1120)
`18.9
`17.8
`18.4
`
`P-value(1)
`0.004
`0.010
`0.007
`
`ACZONE
`(N=1044)
`21.1
`19.8
`20.5
`
`3.2.5 Secondary Efficacy Results
`
`Table 12 presents the results for the secondary efficacy endpoints at Week 12 in both trials. For
`the secondary efficacy endpoints based on lesion counts (i.e., absolute change from baseline in
`total lesion counts and percent change in lesions counts (inflammatory, non-inflammatory, and
`total)), ACZONE gel, 7.5% was statistically superior (p-values ≤ 0.001) to vehicle gel in both
`trials. For the proportion of subjects with an ASIS score of “very good” or “excellent”,
`ACZONE gel, 7.5% was statistically superior to vehicle gel (24% vs. 19%; p-value = 0.015) in
`Trial 006; however, ACZONE gel, 7.5% was not statistically superior to vehicle gel (24% vs.
`22%; p-value = 0.252) in Trial 007. ACZONE gel, 7.5% was not statistically superior to vehicle
`gel for all other secondary efficacy endpoints based on the ASIS.
`
`Reference ID: 3873729
`
`12
`
`

`

`Table 12: Results for the Secondary Efficacy Endpoints at Week 12 (MI(1), LOCF(2), ITT)
`Trial 006(3)
`Trial 007
`
`Vehicle
`Vehicle
`(N=1058)
`(N=1120)
`
`ACZONE
`(N=1044)
`
`P-value
`
`ACZONE
`(N=1118)
`
`P-value
`
`
`
`<0.001(4)
`
`<0.001(4)
`
`<0.001(4)
`
`0.001(4)
`
`0.252(6)
`
`0.057(7)
`
`0.711(6)
`
`0.647(6)
`
`Endpoint
`Absolute Change in
`Total Lesion Counts:
` Mean*
` LS Mean(4)
`Percent Change in
`Total Lesion Counts:
` Mean*
` LS Mean(4)
`Percent Change in
`Inflammatory Lesion
`Counts:
` Mean*
` LS Mean(4)
`Percent Change in
`Non-Inflammatory
`Lesion Counts:
` Mean*
` LS Mean(4)
`ASIS Sign Domain(5):
`Very Good or Excellent
`Absolute Change in
`ASIS Sign Domain:
` Mean
`ASIS Item 1:
`1-grade improvement
`ASIS Item 8:
`561 (53%)
`580 (56%)
`1-grade improvement
`Source: Reviewer’s Analysis (same results as Applicant’s Analysis)
`*The values displayed are the averages over the 20 imputed datasets (MI).
`(1) Missing data for lesion count endpoints were imputed using multiple imputation (MI).
`(2) Missing data for ASIS endpoints were