`
`CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`APPLICATION NUMBER:
`204592Orig1s000
`
`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 TAT I S T I C A L R E V I E W A N D E VA L U AT I O N
`CLINICAL STUDIES
`
`NDA/BLA #:
`Drug Name:
`Indication(s):
`Applicant:
`Date(s):
`Review Priority:
`
`
`
`NDA 204-592
`
`Zorvolex (diclofenac acid)
`
`Treatment of mild to moderate acute pain
`
`Iroko Pharmaceuticals, LLC
`
`Letter date: December 21, 2012, PDUFA date: October 20 , 2013
`
`Standard
`
`II
`
`Feng Li, Ph.D.
`
`Biometrics Division:
`Statistical Reviewer:
`Concurring Reviewers:
`
`Medical Division:
`Clinical Team:
`Project Manager:
`
`
`
`Keywords: NDA review, Clinical Studies
`
`Janice Derr, Ph.D.
`
`Division of Anesthesia, Analgesia, and Addiction Products
`
`Steven Galati, M.D.
`
`Swati Patwardhan, Ph.D.
`
`
`
`
`
`
`
`Reference ID: 3371544
`
`
`
`

`

`
`
`Table of Contents
`
`LIST OF TABLES ....................................................................................................................................................... 3
`
`LIST OF FIGURES ..................................................................................................................................................... 4
`
`1. EXECUTIVE SUMMARY ................................................................................................................................. 5
`
`INTRODUCTION ............................................................................................................................................... 5
`2.
`2.1
`OVERVIEW ...................................................................................................................................................... 6
`2.2
`DATA SOURCES .............................................................................................................................................. 6
`3. 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, Demographic and Baseline Characteristics.......................................................... 8
`3.2.4
`Results and Conclusions ........................................................................................................................ 9
`3.3 EVALUATION OF SAFETY .................................................................................................................................... 14
`4.
` FINDINGS IN SPECIAL/SUBGROUP POPULATIONS ............................................................................ 14
`4.1
`GENDER, AGE AND RACE ............................................................................................................................. 14
`4.2
`OTHER SPECIAL/SUBGROUP POPULATIONS .................................................................................................. 14
`5. SUMMARY AND CONCLUSIONS ................................................................................................................ 15
`5.1
`STATISTICAL ISSUES ..................................................................................................................................... 15
`5.2
`COLLECTIVE EVIDENCE ................................................................................................................................ 15
`5.3
`CONCLUSIONS AND RECOMMENDATIONS ..................................................................................................... 15
`5.4
`LABELING RECOMMENDATIONS ................................................................................................................... 16
`
`
`
`
`
`
`Reference ID: 3371544
`
`2
`
`

`

`
`LIST OF TABLES
`
`Table 1: Subject Disposition − Number (%) of Patients ................................................................................................ 8
`Table 2: Summary of Demographics and Baseline Characteristics ............................................................................... 9
`Table 3: Primary Efficacy Analysis ............................................................................................................................... 9
`Table 4: Rescue Medication − Number (%) of Subjects ............................................................................................. 10
`Table 5: Additional Efficacy Analysis for VASSPID48 – No Imputation After Rescue ............................................ 11
`Table 6: Additional Efficacy Analysis for VASSPID48 – Pre-rescue Score Carried Froward for 6 hours ................. 13
`Table 7: Subgroup Summaries of Primary Endpoint ................................................................................................... 14
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Reference ID: 3371544
`
`3
`
`

`

`
`LIST OF FIGURES
`
`Figure 1: Average Pain Over Time – BOCF After Rescue .......................................................................................... 10
`Figure 2: Average Pain Over Time – No Imputation After Rescue ............................................................................. 11
`Figure 3: Percentage of Subjects with Different Frequency of Rescue Use ................................................................ 12
`Figure 4: Cumulative Distribution of Time to Onset of Analgesia .............................................................................. 13
`
`
`
`
`
`
`
`
`
`
`Reference ID: 3371544
`
`4
`
`

`

`EXECUTIVE SUMMARY
`
`
`1.
`
`
`Iroko Pharmaceuticals, LLC submitted a New Drug Application for Zorvolex, a new formulation
`of diclofenac, seeking an indication for the treatment of mild to moderate acute pain. A Phase 3
`efficacy study in patients with acute pain after bunionectomy was submitted to support the
`efficacy of Zorvolex. Based on my review, the study provided evidence that both Zorvolex
`35 mg and 18 mg three times daily have an analgesic effect in the desired indication in
`comparison to placebo.
`
`The clinical development program of Zorvolex was discussed at several meetings. In January
`2010, the Division of Anesthesia, Analgesia, and Addiction Products (DAAAP) issued an
`agreement letter for the Special Protocol Assessment (SPA) on the Phase 3 efficacy study DIC3-
`08-04. At the End-of-Phase 2 meeting, the applicant was informed that reliance on prior findings
`of efficacy for another diclofenac product and the Study DIC3-08-04 may be adequate to support
`an efficacy claim for the proposed indication.
`
`Study DIC3-08-04 was a randomized, double-blind, multicenter, parallel, active and placebo-
`controlled study evaluating the efficacy and safety of two dosing regimens of Zorvolex in
`subjects with acute postoperative pain after bunionectomy. A total of 428 subjects were
`randomized equally to one of the four treatments for 48 hours: Zorvolex 35 mg three times daily
`(TID), Zorvolex 18 mg TID, celecoxib capsules 200 mg twice daily (BID), or placebo. One
`tablet of hydrocodone/acetaminophen 10 mg/325 mg was permitted every 4 to 6 hours as the
`rescue medication. The primary efficacy variable was the time-weighted sum of pain intensity
`difference from baseline over 48 hours after the first dose. The primary efficacy variable was
`analyzed using an analysis of covariance model with baseline pain score as a covariate and
`treatment as a factor. To control multiplicity, a sequential testing procedure was applied for the
`comparisons of the two doses of Zorvolex with placebo.
`
`The study demonstrated the superiority of both Zorvolex 35 mg TID and 18 mg TID over
`placebo in pain intensity reduction. However, a high percentage of subjects in each treatment
`group took rescue medication at least once during the study. Approximately 82% of the subjects
`in the Zorvolex 35 mg group, 85% of the subjects in the Zorvolex 18 mg group, 85% of the
`subjects in celecoxib group and 97% of the subjects in the placebo group took rescue medication
`for pain management during the study. This might be due to the slow onset of analgesic action.
`For all the treatment groups including the active control, onset of analgesia occurred in less than
`40% of the subjects. Approximately only 31% of the subjects in the Zorvolex 35 mg group, 25%
`of the subjects in the celecoxib group, 24% of the subjects in the Zorvolex 18 mg group, and
`18% of the subjects in the placebo group experienced onset of analgesia within 1 hour after the
`first dose.
`
`
`2.
`
`
`
`
`5
`
`INTRODUCTION
`
`Reference ID: 3371544
`
`

`

`2.1 Overview
`
`
`Diclofenac is an approved drug in the United States as a treatment for multiple indications
`including mild to moderate pain. Iroko Pharmaceuticals, LLC (Iroko) is developing Zorvolex, a
`new formulation of diclofenac in the acid form, for the treatment of mild to moderate acute pain.
`Zorvolex is being filed as a 505(b)(2) application and relies on the previous findings of Cataflam
`safety and efficacy. The applicant believes that the new formulation improves the dissolution and
`absorption of diclofenac. The applicant purports that Zorvolex provides comparable pain relief to
`Cataflam 50 mg at a 20% lower dose of diclofenac and thus potentially leads to an improved
`safety profile.
`
`The clinical development program of Zorvolex capsule was discussed at several meetings under
`IND 103,880. In December 2009, the applicant submitted a Special Protocol Assessment (SPA)
`for the Phase 3 study DIC3-08-04. In January 2010, the division issued an agreement letter on
`the Phase 3 efficacy study. In the agreement letter, the division stated that the secondary
`outcome variables for which no adjustment for multiplicity is planned are considered exploratory
`and therefore would not support a label claim. In addition, the applicant was advised that any
`comparative claims will require replicated demonstration of superiority over the same
`comparator. During the clinical development, the applicant referred to the new formulation as
`. At the End-of-Phase 2 meeting in December 2010, the division informed the
`applicant that the drug product does not meet the agency’s definition of a
`.
`Furthermore, the division stated that reliance on prior findings of efficacy for another diclofenac
`product for acute pain and the proposed Phase 3 efficacy study DIC3-08-04 for the treatment of
`acute postoperative pain after bunionectomy may be adequate to support an efficacy claim for
`this indication.
`
`
`
`2.2 Data Sources
`
`The statistical review is based on data submitted for Study DIC3-08-04. The study data can be
`found at \\Cdsesub1\evsprod\NDA204592\0000\m5\datasets\dic3-08-04.
`
`
`3.
`
`
`STATISTICAL EVALUATION
`
`3.1 Data and Analysis Quality
`
`The applicant submitted study tabulation datasets SDTM and analysis datasets AdaM in CDISC
`format. The submitted datasets and define documents are of acceptable quality.
`
`
`
`3.2 Evaluation of Efficacy
`
`3.2.1 Study Design and Endpoints
`
`
`
`
`Reference ID: 3371544
`
`6
`
`(b) (4)
`
`(b) (4)
`
`

`

`Study DIC3-08-04 was a Phase 3, randomized, double-blind, multicenter, parallel, active and
`placebo-controlled study evaluating the efficacy and safety of two dosing regimens of Zorvolex
`capsules in subjects with acute postoperative pain after bunionectomy. Subject who had a pain
`intensity rating ≥ 40 mm on a 100-mm Visual Analog Scale (VAS) within 9 hours of
`discontinuation of regional anesthesia were to be eligible. Eligible subjects were randomized
`equally to one of the four treatments: Zorvolex 35 mg TID, Zorvolex 18 mg TID, celecoxib
`capsules 200 mg BID, or placebo. Study drug was administered for 48 hours after the first dose.
`
`Pain intensity and pain relief assessments were recorded in the inpatient subject diary at
`scheduled times during the 48-hour period after the first dose (15, 30, and 45 minutes and 1, 1.5,
`2, 3, 4, 5, 6, 7, 8, 12, 16, 20, 24, 32, 40, and 48 hours), and immediately before the first use of
`rescue analgesia if before the 8-hour time point. Pain intensity was assessed based on 100-mm
`Visual Analog Scale (VAS). Pain relief was assessed based on a 5-point categorical scale. Time
`to perceptible and meaningful pain relief was evaluated using the 2-stopwatch method. Time to
`onset of analgesia was measured as time to perceptible pain relief confirmed by meaningful pain
`relief. Pain intensity and pain relief assessments were also recorded before premature study
`termination. Subjects completed a patient’s global evaluation of study drug at the end of the
`treatment period (Day 3) or immediately before the first dose of rescue medication (whichever
`occurs first).
`
`One tablet of hydrocodone/acetaminophen 10 mg/325 mg was permitted as the rescue
`medication. If subjects were unable to tolerate hydrocodone/acetaminophen 10 mg/325 mg, then
`one tablet of oxycodone/acetaminophen 7.5 mg/325 mg was permitted orally every 6 hours as
`needed for pain management. The total daily dosage of rescue medication was not to exceed 6
`tablets.
`
`The primary efficacy variable was the time-weighted sum of pain intensity difference from
`baseline over 48 hours after the first dose (VASSPID48). Secondary efficacy endpoints included
`the VASSPID24, time to onset of analgesia (measured as time to perceptible pain relief
`confirmed by meaningful pain relief), time to rescue, and total amount of rescue.
`
`
`3.2.2 Statistical Methodologies
`
`The primary efficacy variable was analyzed using an analysis of covariance (ANCOVA) model
`with baseline pain score as a covariate and treatment as a factor. The primary analysis population
`included all subjects who were randomized and received at least one dose of study medication.
`To control multiplicity, a sequential testing procedure was carried out for the comparisons of the
`two doses of diclofenac with placebo. Zorvolex 35 mg was compared to placebo first. Zorvolex
`18 mg was compared to placebo only if Zorvolex 35 mg was significantly better than placebo.
`There were no comparisons between Zorvolex and celecoxib in the primary analysis.
`
`Missing pain assessments for subjects who discontinued early due to lack of efficacy or adverse
`events were imputed using a baseline observation carried forward approach (BOCF). Missing
`pain assessments due to other reasons were imputed using a last observation carried forward
`approach (LOCF). For subjects who took any dose of rescue medication, subsequent pain
`
`
`7
`
`Reference ID: 3371544
`
`

`

`assessments after the first dose of rescue medication were disregarded and imputed using a
`BOCF approach. Intermittent missing pain assessments were imputed using linear interpolation.
`
`The applicant conducted a sensitivity analysis for the primary efficacy analysis by adding gender
`as a factor into the ANCOVA model. Sensitivity analyses for the methods to handling missing
`values were not conducted.
`
`Time to onset of analgesia was right censored at 8 hours for subjects who did not experience
`both perceptible pain relief and meaningful pain relief during the 48-hour interval or who
`required rescue medication prior to achieving perceptible or meaningful pain relief.
`
`
`3.2.3 Patient Disposition, Demographic and Baseline Characteristics
`
` A
`
` total of 428 subjects were randomized. All randomized subjects received the study
`medications. The subject disposition is shown in Table 1 with percentages based on the number
`of randomized subjects. Overall, a total of 421 (98%) subjects completed the study. No subjects
`in the Zorvolex 18 mg group discontinued the study early. Both the Zorvolex 35 mg and the
`celecoxib groups had one subject discontinued early due to investigator decision and subject
`request, respectively. Five subjects (5%) in the placebo group discontinued the study early, three
`of which due to lack of efficacy.
`
`
`
`Table 1: Subject Disposition − Number (%) of Patients
`
`
`
`Zorvolex
`35 mg TID
`107
`106 (99%)
`1 (1%)
`
`
`
`1 (1%)
`
`
`
`18 mg TID
`109
`109 (100%)
`0
`
`
`
`
`
`
`
`
`
`Celecoxib BID Placebo
`106
`106
`105 (99%)
`101 (95%)
`1 (1%)
`5 (5%)
`
`
`
`
`
`1 (1%)
`
`
`1 (1%)
`3 (3%)
`
`1 (1%)
`
`
`
`Randomized
`Completed
`Discontinued
`
`Reason for discontinuation
` Subject request
` Investigator decision
` Lack of efficacy
` Lost to follow-up
` Source: Clinical study report (Table 14.1.1)
`
`
`The demographic and baseline characteristics were generally comparable across treatment
`groups. A summary of selected demographic and baseline characteristics is provided in Table 2.
`The summary for race was reproduced using the applicant’s dataset, which differed slightly from
`the clinical study report. The majority of the subjects were female and white. Overall, the mean
`age was about 40 years. Approximately 85% of the subjects were white.
`
`
`
`
`
`
`Reference ID: 3371544
`
`8
`
`

`

`
`
`
`
`Table 2: Summary of Demographics and Baseline Characteristics
`
`
`
`Zorvolex
`
`Celecoxib BID
`18 mg TID
`35 mg TID
`
`N=106
`N=109
`N=107
`Mean age (SD)
`40 (12)
`39 (12)
`39 (12)
`Mean weight (SD) (kg)
`72 (16)
`75 (17)
`77 (19)
`Mean height (SD) (cm)
`165 (8)
`167 (9)
`167 (9)
`Mean BMI (SD) (kg/m2)
`26 (5)
`27 (5)
`27 (6)
`Baseline pain - mean (SD)
`74 (17)
`77 (16)
`74 (16)
` - (Min, Max)
`(40, 100)
`(41, 100)
`(44, 100)
`Gender, n (%)
`
`
`
` Male
`10 (9%)
`15 (14%)
`18 (17%)
` Female
`96 (91%)
`94 (86%)
`89 (83%)
`
`
`
`
`Ethnicity, n (%)
`
`
`
` Hispanic or Latino
`18 (17%)
`24 (22%)
`19 (18%)
` Not Hispanic or Latino
`88 (83%)
`85 (78%)
`88 (82%)
`
`
`
`
`Race, n(%)
`
`
`
` Black or African American
`22 (21%)
`19 (17%)
`17 (16%)
` White or Caucasian
`73 (69%)
`86 (79%)
`84 (78%)
` Other
`11 (10%)
`4 (4%)
`6 (6%)
` Source: Clinical study report (Table 14.1.2) and applicant’s datasets; SD: standard deviation.
`
`
`
`Placebo
`N=106
`40 (13)
`73 (14)
`167 (8)
`26 (5)
`76 (16)
`(40, 100)
`
`14 (13%)
`92 (87%)
`
`
`17 (16%)
`89 (84%)
`
`
`19 (18%)
`79 (74%)
`8 (8%)
`
`3.2.4 Results and Conclusions
`
`I replicated the applicant’s results for the primary efficacy analysis. Table 3 shows the results
`from the primary efficacy analysis. Both Zorvolex 35 mg and 18 mg were superior to placebo in
`terms of the primary efficacy endpoint. The applicant conducted a sensitivity analysis by adding
`gender as an extra factor into the model. The results of this sensitivity analysis were similar to
`those of the primary analysis.
`
`
`
`Table 3: Primary Efficacy Analysis
`
`
`
`
`Zorvolex
`
`Placebo
`Celecoxib BID
`18 mg TID
`35 mg TID
`Statistics
`106
`106
`109
`107
`n
`77 (86)
`390 (86)
`393 (85)
`524 (86)
`LS Mean (SE)
`( -93, 247)
`(220, 560)
`(225, 561)
`(355, 693)
`95% CI
`
`
`
`
`
`
`313 (122)
`316 (121)
`447 (122)
`Difference in LS mean(SE)
`
`(72, 554)
`(77, 555)
`(207, 687)
`95% CI for diff. in LS mean
`
`0.01
`0.01
`<0.001
`p-value for treatment effect
` Source: Clinical study report (Table 14.2.1.1); SE: standard error; CI: confidence interval; LS: least square.
`
`
`
`Reference ID: 3371544
`
`9
`
`

`

`
`The applicant used a hybrid LOCF/BOCF approach to impute pain scores after early
`discontinuation. Although it is a single imputation method, I am not concerned about it in this
`study as very few subjects (2%) discontinued early. The applicant replaced all the pain scores
`after the first use of the rescue medication with the baseline observations. When the percentage
`of subjects who take rescue medications is high, the approach to handling the pain scores after
`rescue may substantially influence the comparisons among treatments. In the acute pain setting,
`it is often likely that subjects will take rescue medications. Table 4 presents the percentage of
`subjects who took rescue medications for pain management during the study. For all treatment
`groups, more than 80% of the subjects took rescue. Placebo group had the highest percentage of
`subjects who took rescue. The majority of the subjects took their first rescue within 8 hours after
`the first dose, that is, before the second dose of the study medication. Figure 1 depicts the
`average pain intensity over time for each treatment group during the 48 hours after the first dose
`with pain scores after rescue imputed using a BOCF approach. Although the placebo group had
`the worst pain on average, there was not much pain reduction for all treatment groups. All the
`pain curves are rather flat after 10 hours. This is because the majority of the subjects took rescue
`during the first 8 hours and their pain scores after the first rescue were replaced by the
`corresponding baseline values.
`
`
`Table 4: Rescue Medication − Number (%) of Subjects
`
`Zorvolex
`35 mg TID
`107
`88 (82%)
`84 (78%)
`
`
`18 mg TID Celecoxib BID
`106
`109
`90 (85%)
`93 (85%)
`87 (82%)
`88 (81%)
`
`
`Placebo
`106
`103 (97%)
`101 (95%)
`
`Figure 1: Average Pain Over Time – BOCF After Rescue
`
`
`
`
`
`Randomized
`Subjects who took rescue
`Subjects who took rescue within 8 hours after first dose
`Source: applicant’s datasets
`
`
`
`
`
`Reference ID: 3371544
`
`
`
`10
`
`

`

`Figure 2 displays the average pain intensity over time for each treatment group without
`imputation for taking rescue, that is, the actual observed pain scores after rescue were used. The
`overall trend of actual pain reduction over time is apparent for each treatment group. Among the
`treatments, subjects in the placebo group experienced the least pain reduction. The separation of
`the pain curve of the placebo from the three active treatments occurred after approximately 3
`hours after dosing. To compare the treatment effects under the influence of rescue medications in
`terms of the primary efficacy endpoint, I calculated the summed pain intensity difference over 48
`hours using the observed pain scores after rescue and conducted an analysis using the same
`ANCOVA model as the one used in the primary analysis. The analysis results are presented in
`Table 5. The differences between the three active treatments and placebo were all statistically
`significant, which indicates that the active treatments in combination with the rescue medications
`produced superior analgesic effects to placebo in combination with the rescue medications.
`
`
`Figure 2: Average Pain Over Time – No Imputation After Rescue
`
`
`Table 5: Additional Efficacy Analysis for VASSPID48 – No Imputation After Rescue
`
`
`
`
`
`
`Celecoxib BID
`106
`2159 (82)
`(1997,2320)
`
`498 (116)
`(269,726)
`<0.0001
`
`
`Placebo
`106
`1661 (82)
`(1499,1823)
`
`
`
`
`
`11
`
`
`
`
`
`Zorvolex
`
`18 mg TID
`35 mg TID
`Statistics
`109
`107
`n
`2293 (81)
`2392 (82)
`LS Mean (SE)
`(2134,2452)
`(2232,2553)
`95% CI
`
`
`
`632 (115)
`731 (116)
`Difference in LS mean(SE)
`(405,859)
`(503,960)
`95% CI for diff. in LS mean
`<0.0001
`<0.0001
`p-value for treatment effect
` SE: standard error; CI: confidence interval; LS: least square.
`
`
`
`Reference ID: 3371544
`
`

`

`
`To further compare the usage of rescue medication, I depicted the percentages of subjects who
`took rescue medications for pain control over different frequencies of rescue administration for
`each treatment group in Figure 3. There was consistently higher percentage of subjects in the
`placebo group than in the active treatment groups for each category of frequency. For example,
`approximately 85% of the subjects in the placebo group took rescue at least 3 times. In contrast,
`about 40% of the subjects in the Zorvolex 35 mg, 50% in the Zorvolex 18 mg, and 55% in the
`celecoxib groups took rescue medication at least 3 times.
`
`
`
`
`Figure 3: Percentage of Subjects with Different Frequency of Rescue Use
`
`
`
`
`For handling pain scores after taking rescue medications, DAAAP currently recommends
`replacing the pain scores that might be affected by the rescue medications with the pre-rescue
`pain score for efficacy analyses to minimize the impact of rescue medication. Thus, I conducted
`additional analyses in which the pain scores within a time window after taking a rescue
`medication were replaced with the pre-rescue pain score. Table 6 presents the results from an
`analysis in which the pain scores within 6 hours after rescue were replaced with the pre-rescue
`score. The analyses results were in favor of the active treatments. The pain curves showed the
`similar pattern as that observed in Figure 2. My analyses using different lengths of time window
`(such as 4 or 8 hours) yielded similar results.
`
`
`
`
`
`
`
`
`Reference ID: 3371544
`
`12
`
`

`

`Table 6: Additional Efficacy Analysis for VASSPID48 — Pre-rescue Score Carried Froward for 6 hours
`
` Zorvolex
`
`Statistics
`n
`
`LS Mean (SE)
`95% CI
`
`35 mg T11)
`107
`
`2221(82)
`(2060,2383)
`
`18 ES TlD
`109
`
`2097(81)
`(1937,2257)
`
`Celecoxib BID
`106
`
`Placebo
`106
`
`1976(83)
`(1813,2138)
`
`1407(83)
`(1245,1569)
`
`690(1 16)
`814(11 7)
`Difference in LS mean(SE)
`(462.918)
`(585.1043)
`95% CI for diff. in LS mean
`<0.0001
`<0.0001
`p-value for treatment effect
`SE: standard error; CI: confidence interval; LS: least square.
`
`569(1 1 7)
`(339.798)
`<0.0001
`
`The cumulative distribution of time to onset of analgesia for each treatment group is shown in
`Figure 4. For all the treatment groups, onset of analgesia occurred in less than 40% of the
`subjects. Among the treatment groups, the Zorvolex 35 mg group had the highest percentage of
`subjects who experienced onset of analgesia whereas the placebo group had the lowest
`percentage. Approximately 31% of the subjects in the Zorvolex 35 mg group, 25% of the
`subjects in the celecoxib group, 24% of the subjects in the Zorvolex 18 mg group, and 18% of
`the subjects in the placebo group experienced onset of analgesia within 1 hour after the first
`dose.
`
`Figure 4: Cumulative Distribution of Time to Onset of Analgesia
`
`. ;‘ acebu Percent
`
`
`:t‘ButjectzDyermm'irqAnalgesmll:
`
`— “r' 0551:: (ll) (“cumin-.31 Swazi: 35 sq II:
`- - — :acloz'enc
`fumtzoa Capsule LE tq II:
`— — Celecanb Capsule :32 mg 312 [ill In; 3.5.1an Bosch
`
`Source: Clinical Study Report (Figure 14.2.4)
`
`T22 tEouxs
`
`Analyses results of the secondary efficacy endpoints were supportive to the primary analysis.
`
`13
`
`Reference ID: 3371544
`
`

`

`3.3 Evaluation of Safety
`
`
`
`
`The evaluation of the safety data was conducted by Dr. Steven Galati. The reader is referred to
`Dr. Galati’s review for detailed information regarding the adverse event profile.
`
`
`4.
`
` FINDINGS IN SPECIAL/SUBGROUP POPULATIONS
`
`
`The applicant performed the subgroup analyses of the primary endpoint by age, race and gender
`using the per-protocol population (PP). I conducted subgroup summaries by gender, age, and
`race for all the randomized subjects. For age, subjects were classified as ≤ 45 or >45 years old.
`For race, subjects were categorized as black, white or other races. The findings from the
`subgroups summaries were consistent with those observed in the overall population. All the
`active treatment groups were numerically better than the placebo group in the subpopulations.
`
`
`
`4.1 Gender, Age and Race
`
`Table 7 shows the subgroup summaries for gender, age and race.
`
`
`
`
`
`
`
`
`Subgroups
`Sex
` Female
`
` Male
`
`Table 7: Subgroup Summaries of Primary Endpoint
`
`
`
`Statistics
`
`n (%)
`Mean (SD)
`n (%)
`Mean (SD)
`
`n (%)
`Mean (SD)
`n (%)
`Mean (SD)
`
`n (%)
`Mean (SD)
`n (%)
`Mean (SD)
`n (%)
`Mean (SD)
`
`Zorvolex
`35 mg TID
`(N=107)
`
`89 (83%)
`479 (1109)
`18 (17%)
`751 (1326)
`
`71 (66%)
`415 (970)
`36 (34%)
`739 (1423)
`
`17 (16%)
`1120 (1658)
`84 (79%)
`399 (961)
`6 (6%)
`596 (1490)
`
`18 mg TID
`(N=109)
`
`94 (86%)
`375 (922)
`15 (14%)
`505 (1057)
`
`74 (68%)
`346 (921)
`35 (32%)
`492 (977)
`
`19 (17%)
`533 (1075)
`86 (79%)
`343 (876)
`4 (4%)
`808 (1583)
`
`
`Celecoxib BID
`(N=106)
`
`96 (91%)
`353 (888)
`10 (9%)
`746 (1228)
`
`71 (67%)
`254 (746)
`35 (33%)
`667 (1174)
`
`22 (21%)
`120 (517)
`73 (69%)
`477 (1025)
`11 (10%)
`356 (799)
`
`
`Placebo
`(N=106)
`
`92 (87%)
`77 (350)
`14 (13%)
`79 (286)
`
`68 (64%)
`34 (152)
`38 (36%)
`154 (527)
`
`19 (18%)
`60 (246)
`79 (75%)
`87 (376)
`8 (8%)
`20 (70)
`
`
`
`
`
`Age
` <=45
`
` >45
`
`Race
` Black or African American
`
` White or Caucasian
`
` Other
`
`
`
`
`
`
`4.2 Other Special/Subgroup Populations
`
`No other subgroup summaries were performed.
`
`
`
`
`14
`
`Reference ID: 3371544
`
`

`

`
`SUMMARY AND CONCLUSIONS
`
`
`5.
`
`
`5.1 Statistical Issues
`
`
`The applicant used a hybrid LOCF/BOCF method to impute pain scores after early
`discontinuation. In 2010, the National Academy of Science (NAS) released a report on missing
`values. The report does not recommend single imputation approaches for imputing missing
`values due to dropouts for reasons including the difficulty to justify the underlying assumptions
`and the underestimation of the uncertainty of missing observation. Although the proposed
`LOCF/BOCF method is a single imputation method, I am not concerned about it in this study as
`very few subjects (2%) discontinued early.
`
`To account for the influence of rescue medication, the applicant replaced all pain scores after the
`first use of the rescue medication with the baseline values. In this study, more than 80% of the
`subjects took rescue and the majority of them used rescue within several hours after the first
`dose. Thus, I was initially concerned about the applicant’s method for handling the pain scores
`after rescue as the difference between treatments in terms of the primary endpoints might be
`purely driven by the differential percentages of subjects who took rescue. My concern was
`alleviated after my sensitivity analyses including the pain scores after rescue or replacing the
`pain scores affected by rescue with the pre-rescue pain score also yielded statistically significant
`results.
`
`
`
`5.2 Collective Evidence
`
`There was statistically significant difference between each dose of Zorvolex and placebo in
`terms of the sum of pain intensity difference over 48 hours. The results were not sensitive to the
`methods for handling missing values or the pain scores after rescue use. The secondary endpoints
`were also consistently in favor of Zorvolex in comparison to placebo. All the treatment groups
`had high percentages of subjects who took rescue for pain control. The Zorvolex groups had
`similar percentages of subjects who used rescue as the active control, celecoxib. The placebo
`group used more rescue than the active treatment group.
`
`
`
`5.3 Conclusions and Recommendations
`
`The study has demonstrated that both Zorvolex 35 mg and 18 mg were more efficacious than
`placebo in acute pain reduction. The review team will need to consider the totality of evidence
`including findings from clinical pharmacology to decide whether the benefit-risk profile justify
`the approval of the product. Since high percentage of subjects also took rescue medication for
`pain management during the study, I would recommend the applicant include the information
`about the rescue medication and the percentage of subjects who used rescue in the clinical study
`section of the label if the division decides to approve the product.
`
`
`
`15
`
`Reference ID: 3371544
`
`

`

`
`
`
`5.4 Labeling Recommendations
`
`
`The applicant submitted the following wording for the clinical study section of the label for
`review:
`
`
`
`
`Reference ID: 3371544
`
`16
`
`(b) (4)
`
`

`

`
`
`17
`
`
`
`Reference ID: 3371544
`Reference ID: 3371544
`
`(b) (4)
`
`

`

`---------------------------------------------------------------------------------------------------------
`This is a representation of an electronic record that was signed
`electronically and this page is the manifestation of the electronic
`signature.
`---------------------------------------------------------------------------------------------------------
`/s/
`----------------------------------------------------
`
`FENG LI
`09/11/2013
`
`JANICE A DERR
`09/11/2013
`
`Reference ID: 3371544
`
`

`

`STATISTICS FILING CHECKLIST FOR NDA 204-592
`
`Applicant: Iroko Pharmaceuticals Stamp Date: Dec 20, 2012
`NDA/BLA Type: 505(b)(2)
`
`
`
`
`NDA