CENTER FOR DRUG EVALUATION AND
`
`RESEARCH
`
`APPLICA TION NUMBER:
`
`21-450
`
`STATISTICAL REVIEW! S!
`
`
`
`

`

`v
`
`:9— vim 4...... e...
`
`
`
` Anselm ' STATISTICAL REVIEW AND EVALUATION
`
`
`
`DEPARTMENT OF IEEALTH AND HUMAN SERVICES
`PUBLIC HEALTH SERVICE
`FOOD AND DRUG ADMINISTRATION
`CENTER FOR DRUG EVALUATION AND RESEARCH
`
`Neumpharmacological Drug Products (HFD-lZO)
`MEDICAL DIVISION:
`BIOMETRICS DIVISION: Division of Biometrics I (HID-710)
`
`NDA NUMBER:
`
`NDA 21-450
`
`DRUG NAME:
`
`Zomig (zolmitriptan) Nasal Spray
`
`INDICATION:
`
`Treatment of migraine with/without aura in adults
`
`SPONSOR:
`
`AstraZeneca Pharmaceuticals LP
`
`Cover letter and documents (CDER REC’D
`DOCUMENTS REVIEWED:
`Dates: 27-Mar-2003, 17-Apr-2003 and 25-Apr-2003) including SAS data base
`
`xz-test, logistic regression, odds ratio,
`STATISTICAL KEY WORDS:
`Bonferroni adjustment, spending function, interim analysis
`
`STATISTICAL REVIEWER:
`
`Yong-Cheng Wang, PhD. (RFD—710)
`
`STATISTICAL TEAM LEADER: Kun Jin, PhD. (FIFE-710)
`
`DBI DEPUTY DIRECTOR:
`
`George Chi, PhD. (HFD—7 10)
`
`CLINICAL REVIEWER:
`
`Kevin A. Prohaska D.O., MD. (HFD-IZO)
`
`CLINICAL TEAM LEADER:
`
`Armando Oliva, MD. (HFD—IZO)
`
`PROJECT MANAGER:
`
`Lana Chen (HFD-IZO)
`
`Distribution: NDA 21-450
`HFD-lZO/Dr. Katz
`HFD-lZO/Dr. Oliva
`HFD-lZO/Dr. Prohaska
`
`HFD-lZO/Ms. Chen
`HFD-700/Dr. Anello
`HFD-7lO/Dr. Chi
`I-IFD-710/Dr. Jin
`
`RFD-7 lO/Dr. Wang
`
`File and Date: C:\NDA\Zomig Nasal\ReV'iew\Zolmitriptan Nasal.doc
`
`7-16—2003
`
`
`
`

`

`
`
`Table of Contents
`
`1
`
`2
`
`EXECUTIVE SUMMARY OF STATISTICAL FINDINGS .................... l
`
`1.1 RECOMMENDATIONS AND CONCLUSIONS ..................................................... I
`1.2 BRIEF OVERVIEW OF CLINICAL STUDIES ...................................................... 1
`1.3 STATISTICAL ISSUES AND FINDINGS ............................................................. 3
`
`INTRODUCTION ......................................................................................... 4
`
`2.1 OVERVIEW ................................................................................................... 4
`
`2.1.1 Background ......................................................................................... 5
`2.1.2 Major Statistical Issues ........................................................................ 6
`2.2 DATA SOURCES ............................................................................................ 6
`
`3
`
`STATISTICAL EVALUATION .................................................................. 6
`
`3.! EVALUATION OF EFFICACY .......................................................................... 6
`
`3.1.1 StudyJMCH ........................................................................................ 6
`3.1.1.1 Introduction ..................................................................................... 6
`3.1.1.2 Statistical Issues .............................................................................. 7
`
`3.1.1.3 Study Objectives ............................................................................. 7
`3.1.1.4 Efficacy Endpoints .......................................................................... 7
`3.1.1.5 Sample Size Considerations ............................................................ 7
`3.1.1.6 Stratification .................................................................................... 8
`
`3.1.1.7 Interim Analysis .............................................................................. 8
`3.1.1.8 Efficacy Analysis Methods ............................................................. 8
`3.1.1.9 Sponsor’s Results and Reviewer’s Findings/Comments ................. 9
`3.1.1.9.] Baseline Characteristics .......................................................... 9
`
`3.1.1.9.2 Primary Efficacy Analyses .................................................... 10
`3.1.1911 Secondary Efficacy Analyses ................................................ 1 l
`3.1.1.10 Sponsor's Conclusions and Reviewer’s Conclusions/Comments ..... 13
`3.2 EVALUATION OF SAFETY ............................................................................ 13
`
`FINDINGS IN SPECIAUSUBGROUP POPULATIONS ...................... 13
`
`SUMMARY AND CONCLUSIONS .......................................................... 13
`
`4 5
`
`5.1 STATISTICAL ISSUES AND COLLECTIVE EVIDENCE ..................................... 13
`5.2 CONCLUSIONS AND RECOMMENDATIONS ................................................... l4
`
`6
`
`APPENDICES ............................................................................................. 15
`
`APPEARS THIS WAY
`0?! @RlGll‘iAl
`
`
`
`

`

`
`
`1 Executive Summary of Statistical Findings
`
`1.1 Recommendations and Conclusions
`
`Based on the collective evidences and findings, in this statistical reviewer’s
`opinion the interim data and results of study 311CUS/0022 support the sponsor’s
`efficacy claim of ZOMIG® (zolmitriptan) 5 mg nasal with respect to the headache
`response endpoint for the adult patients with migraine, using first attack analysis.
`The data and results of the study show that the primary endpoint, 2 hour headache
`response, is statistically significantly improved in the treatment arm to the placebo
`arm for the intention-to-treat (FIT) population (p-value=0.0006).
`
`There was no statistically significant differences demonstrated for Zolmitriptan
`Nasal Spray dose of 5.0 mg to placebo in the secondary variable of nausea at 2
`hours post dose. There was statistically significant differences demonstrated for
`Zolmitriptan Nasal Spray doses of 5.0 mg to placebo in the secondary variable of
`photophobia at 2 hours post-dose. There was no evidence to demonstrate the
`statistically significant difference for Zolrnitriptan Nasal Spray dOSe of 0.5 mg to
`placebo in phonophobia at 2 hours post-dose.
`
`Those results are consistent with the original NDA submission of Zolmitn'ptan
`Nasal Spray. We therefore recommend Approval for the treatment of adult
`patients with migraine.
`
`1.2 Brief Overview of Clinical Studies
`
`The zolmitriptan nasal spray NDA 21-450 was submitted on February 27, 2002.
`In the action letter dated December 19, 2002, FDA classified the submission as
`‘approvable'. Additional information was requested from AstraZeneca to address
`concerns regarding the clinical efiicacy of the commercial zolmitriptan nasal
`spray device. At a teleconference (February 11, 2003), FDA agreed that the
`provision of efficacy data from an ongoing, placebo-controlled clinical trial using
`the commercial zolmitriptan nasal spray device (31 lCUS/OOZZ) was an
`acceptable approach.
`
`This interim analysis consists of results from a subset of the Study 311CUS/0022
`in 210 adult patients who treated the first headache attack with study medication
`and provided efficacy assessments. This sample size provided adequate power
`(for 2-hour headache response) to show superiority over placebo and confirm the
`clinical efficacy of the zolmitriptan 5-mg commercial nasal spray device.
`
`The primary objective for this interim analysis is to evaluate the efficacy (as
`assessed by the 2 hour headache response) of zoImitriptan 5-mg nasal spray
`compared to placebo in the acute treatment of adult patients with migraine, using
`__ 5—H..-
`
`("“n,
`
`

`

`
`
`STATISTICAL REVIEW AND EVALUATION
`
`first attack analysis. The primary endpoint is the 2-hour headache response of the
`first treated attack.
`
`The sponsor planned an adjustment for Type I error for both the interim and final
`full studies. The alpha Spending fimction methodology based on Hwang, Shih,
`and deCani (1990) 'y-family approach was used to control the overall two-sided
`type I error rate at 5% for both the interim and final analyses. The 2-sided
`significance boundaries for the p-values were calculated and pre-specified to be
`0.0027 and 0.0479 (based on 7= -2 and information fraction of t=15%) for the
`interim and the final analyses, respectively. That is, the statistical significance of
`the analysis results for the primary efficacy parameter of 2-hour headache
`response was tested against significance level of 0.0027 for the interim and
`0.0479 for the final analyses, respectively.
`
`The first 210 adult men and women who treated the first migraine attack with
`study medication, and who had an established diagnosis of migraine headache,
`with or without aura, as defined by IHS criteria were included in this analysis.
`This sample size provided approximately 90% power of showing a difference (at
`0.27% two-sided level of significance) in headache response rate between the
`zolmitriptan 5-mg nasal spray dose and placebo at 2 hours after treatment.
`Calculations were based on the assumption that the headache response rate at 2
`hours would be 39% for placebo and 69% for zoimitriptan S-mg nasal Spray.
`
`Reviewer ’5 Comments:
`
`This statistical reviewer used EaSt 2000 statistical software to verify 0t adjustment
`for the interim and final analyses and summarized the results in Table 1. The
`calculation was based on the overall significance level = 0.05, power = 90%,
`proportion response of control = 6%, and proportion response of treatment = l 1%
`which were used by the sponsor for sample size determination of the full study.
`
`
`0t Ad'ustment for Interim and Final Anal ses -— FDA Anal sis
`Table 1.
`
`
`
`
`This statistical reviewer has the following comments.
`
`

`

`’ f STATISTICAL REVIEW AND EVALUATION "
`
`0 From Table l, the boundaries to reject Ho A z 0.0 and 0.5 are corresponding to
`O’Brien-Fleming and Pocock boundaries, respectively. The 0t=0.0027 and
`0.0479 of interim and final analyses used by the Sponsor is correSponding to a
`boundary which is between 0.3 and 0.4.
`
`0 All requested maximum subjects for A = 0.0 to 0.5 are smaller than n = 1592
`which was planned by the sponsor for the full study.
`
`0 The number of planned looks is 2 which is including the final look.
`
`1.3
`
`Statistical Issues and Findings
`
`Statistical Issues:
`
`0
`
`0
`
`In the primary analysis, the endpoint was defined as 2 hours. If there was a
`missing data at the endpoint, the last observation would be carried forward
`(LOCF), i.e., the last post—treatment measurement was used in the analysis. In
`the confirmatory analysis, missing data at each time point was not imputed
`and the analysis was based on only the subjects who completed
`measurements. The Sponsor did not use LOCF algorithm for the primary
`efficacy analysis. This statistical reviewer used LOCF algorithm to confirm
`the Sponsor’s results.
`In the confirmatory analyses, the secondary efficacy variables: headache
`responses at the time point 1 hour and 4 hours supported the primary efficacy
`result. However, a comparative analysis for the usual migraine-associated
`symptoms of Nausea, Photophobia and Phonophobia did not support the
`primary efficacy analysis. Clinically, those symptoms are highly associated
`with migraine.
`0 There were multiple comparisons in the secondary analyses. The NBA
`submission did not adjust the overall significance level (0t=0.05) for the
`comparisons of secondary endpoints.
`
`
`
`Findings:
`
`Table 2 gives the summary of efficacy results of primary endpoint, the 2—hour
`headache response of the first treated attack, for the ITT population. A total of
`108 patients on the zolmitn'ptan nasal spray 5 mg arm and 102 patients on the
`placebo arm were included in the efficacy analysis. The number of patients with
`headache response at 2-hour were 76 (70.4%) and 48 (47.1%) in the two arms,
`respectively. There was a statistically significant difference (p=0.0006) between
`the two treatment groups.
`
`
`
`

`

`
`
`.
`i‘
`
`-
`
`-_..:,,
`
`I
`
`" STATISTICAL REVIEW AND EVALUATION
`
`Table 2.
`
`Primary endpoint: Headache response at 2 hours -—- first
`attack gzolmitriptan nasal spray vs. placebo!: FDA Analysis
`,
`Zolmitriptan nasal spray 5.0 mg
`Placebo
`
`(N = 102
`(N = 108 '
`100
`108
`Patients evaluated at 2 h (n)
`43 (47-1)
`76 (70.4)
`Patients with headache response at 2 h (‘l/c.)c
`_fi_____..._.____——_____.___m___
`p—value
`0.0006
`Odds ratio°
`2.57
`95% confidence limits‘
`(1.5, 4.9!
`Statistical reviewer’s results based on the analysis data sets provided by the sponsor.
`' Patients received trial medication and treated a migraine attack
`b Two patients from placebo group were not evaluated: One patient had missing data at 2 hour; the other
`patient had taken escape medication before 2 hours when migraine headache pain was mild.
`c Headache response is as a reduction in headache intensity from moderate or severe to mild or none.
`d P-value is calculated by chi—square test.
`‘ Odds ratio and 95% confidence limits (C1) are estimated by SAS FREQ procedure.
`
`_
`
`Reviewer '3 Comments:
`
`The sponsor's efficacy claim was based on the ITT population which was not
`including two patients randomized but not evaluated at 2 hour: One patient had
`missing data at 2 hour and the other patient had taken escape medication before 2
`hours when migraine headache pain was mild. This statistical reviewer used
`LOCF algorithm for the efficacy analysis, which included the two patients.
`
`2
`
`Introduction
`
`2.1 Overview
`
`Zolmitriptan is a potent and selective 5-HTIBHD receptor agonist that has been
`developed for the acute treatment of migraine headache with or without aura. It
`was originally produced in conventional oral tablet form, and in this form was
`approved for use in the acute treatment of migraine in the United States (NBA 20-
`768, approved November 25, 1997) and most of Europe. More recently, an orally
`disintegrating tablet of zolmitriptan (ZOMIG FASTMELT) has been developed
`and is used in some European countries and the United States (ZOMIG—ZMT,
`NDA 21-231, approved on February 13, 2001). This formulation dissolves
`‘
`a rapidly in the mouth and dose not need to be swallowed with water, and may thus '
`often more convenient dosing, particularly in the patients who suffer nausea
`and/or vomiting as part of a migraine attack.
`
`
`
`

`

`
`
` ' STATISTICAL REVIEW AND EVALUATION
`
`2.1.1 Background
`
`Migraine is a debilitating and recurring disease affecting approximately 18.2% of
`women and 6.5% of men in the United States population (Stewart et a1 2001).
`The typical migraine is characterized by a throbbing headache that is usually
`unilateral, made worse by movement, and is often associated with nausea,
`vomiting, photophobia, and phonophobia. Attacks may be preceded by an aura in
`which transient focal neurological symptoms, usually visual disturbances, occur
`(Jensen et a1 1986).
`
`The pharmacotherapy of migraine includes prophylaxis to reduce the frequency
`and severity of attacks, but the incomplete effectiveness and potential side effects
`of prophylactic drugs mean that symptomatic treatment of attacks is the usual
`approach. Symptomatic treatment of migraine has been revolutionized in the last
`decade by the advent of the triptans. Zolmitriptan, a nonsulphonamide synthetic
`analogue of 5-HT acts as a selective agonist at the 5-Hrrartn receptor subtype.
`In
`vitro studies have shown that zolmitriptan has a high affinity for both human 5-
`H111) and 5-Hrus receptors, with approximately a 10-fold selectivity for the former.
`Zolmitriptan is not recognized by other 5-HT receptors, with the exception of a
`modest affinity at the human 5~Hm receptor, and is also devoid of
`pharmacological activity at a wide variety of monoamine receptors (zolmitriptan
`investigator’s brochure [13]).
`
`Peripheral inhibitory effects on the trigeminovascular system, characterized by the
`vasoconsniction of the cranial blood vessels can be observed after the
`
`administration of zolrnitriptan (Plosker 1994). Data from animal models have
`indicated that zolfnitriptan has the ability to inhibit the trigeminovascular system
`at central sites because it can cross the blood-brain barrier and access the 5—Hru)
`
`receptor thereby inhibiting the activated central components of the
`trigeminovascular system within the trigeminal nucleus caudalis (Goadsby and
`Hoskin 1995; Mills et a1 1995). Zolmitriptan is also thought to inhibit
`neuropeptide release after trigeminal activation (Goadsby and Edvinsson 1994).
`
`Most triptans are taken orally, but absorption of some oral triptans is affected by
`the gastric stasis experienced by many migraineurs. Consequently, other, non-
`oral routes of administration are desirable. Compared with the oral route,
`intranasal may permit faster and more complete absorption, and avoid first-pass
`metabolism. These characteristics may be associated with faster onset of action,
`which would be valuable to all migraineurs, not just those who are unable to take
`treatments orally. Intranasal administration is also more acceptable to most
`patients than subcutaneous injection.
`
`Zomig (zolmitriptan), developed as an oral tablet, and as an orally disintegrating
`tablet, was approved by FDA as being safe and effective for the acute treatment of
`
`
`
`

`

` STATISTICAL REVIEW AND EVALUATION
`
`migraine attacks with and without aura in adults. The oral tablet has been
`approved for marketing in more than 75 countries worldwide. A nasal spray
`formulation is currently marketed as the commercial device in Austria, Germany,
`Sweden, and the United Kingdom.
`
`2.1.2 Major Statistical Issues
`
`The major statistical issues can be found in Section 1.3.
`
`2.2 Data Sources
`
`The data sets analyzed were submitted by the sponsor on March 27 and April l7,
`2003. All data sets analyzed are electronic documents and located in the
`Electronic Document Room (EDR) of CDER of FDA under the Letter Date “27-
`MAR-2003” and “17-APR-2003”, respectively. The main data set for the
`efficacy analysis is “DIARYV’” which describes the Diary card.
`
`3 Statistical Evaluation
`
`3.1 Evaluation of Efficacy
`
`This interim analysis was used for efficacy evaluation. The data was from an
`ongoing, placebo-controlled clinical trial using the commercial zolmitriptan nasal
`spray device (311CUSI0022) that FDA agreed to be an acceptable approach as the
`additional information for the approvable zolmitriptan nasal spray NDA 21-450.
`The interim analysis consists of results from a subset of the Study 311CUS/0022
`in 210 adult patients who treated the first headache attack with study medication
`and provided efficacy assessments. All 210 patients consist of the ITT population
`for the primary efficacy analysis.
`
`3.1.1
`
`Interim Analysis of Study 311CUSI0022
`
`3.1.1.1 Introduction
`
`Study 311CUS/0022 was a multi-center, randomized, double-blind, parallel-
`group, placebo-controlled trial to compare the efficacy and tolerability of the
`zolmiuiptan 5-mg nasal spray dose and placebo in the acute treatment of
`migraine. The trial is being conducted at 162 investigation sites in the United
`States. Evaluable patients were those who treated at least 1 migraine headache
`with trial treatment (up to 2 migraine headaches can be treated with study
`medication). Evaluable patients for the interim analysis were the first 210
`patients from 36 centers who treated the first migraine headache and provided
`efficacy assessments.
`
`
`
`
`
`

`

`
`
`STATISTICAL REVIEW AND EVALUATION
`
`3.1.1.2 Statistical Issues
`
`The major statistical issues can be found in Section 1.3.
`
`3.1.1.3 Study Objectives
`
`Full Study Objective: The primary objective of the study was to evaluate
`early efficacy (as assessed by the percentage of responders) of a zohnitriptan
`5-mg nasal spray dose in the acute treatment of adult patients with migraine.
`The secondary objective of this study was to further evaluate the efficacy,
`safety, and tolerability of the zolmitriptan 5-mg nasal spray dose in the acute
`treatment of migraine.
`
`Interim Analysis Objective: The primary objective was to evaluate the
`efficacy (as assessed by the 2 hour headache reSponse) of zolrnitriptan 5-mg
`nasal spray compared to placebo in the acute treatment of adult patients with
`migraine, using first attack analysis. Secondary objectives were to evaluate
`headache response at 15 minutes, 30 minutes, 1 hour, and 4 hours using the
`first attack, and to evaluate the resolution of migraine-associated symptoms
`(nausea, photophobia, and phonophobia) at 2 hours.
`
`_.—-—.
`
`3.1.1.4 Efficacy Endpoints
`
`Full Study Efficacy Endpoints: The primary efficacy endpoints were to be
`headache response at 15 minutes, 30 minutes, 1 hour, and 2 hours after
`initial treatment with trial medication. Headache reSponse is defined as an
`improvement in migraine headache intensity from severe or moderate to
`mild or none. The primary analysis of the efficacy endpoints will be based
`on the first attack data.
`
`Interim Analysis Efficacy Endpoints: The primary efficacy measurement
`was headache response at 2 hours afier initial treatment of the first attack
`with trial medication. Headache response was defined as an improvement in
`migraine headache intensity fi'om severe or moderate to mild or none. In
`addition, the following secondary endpoints were assessed: headache
`response rate at 15 minutes, 30 minutes, 1 hour, and 4 hours after treatment
`resolution of non-headache symptoms of migraine (i.e., nausea,
`photophobia, and phonophobia) at 2 hours after treatment.
`
`3.1.1.5 Sample Size Considerations
`
`Sample Size for Full Study: Approximately 1592 adult men and women
`who had an established diagnosis of migraine headache, with or without
`m. ~...-... .1
`
`

`

`
`
`
`
`aura, as defined by the International Headache Society (1H8) criteria were to
`be enrolled to obtain 1384 evaluable patients (692 patients per treatment
`.
`group). Assuming a 13% rate of withdrawal and unevaluable patients
`(withdrawal rate from 311IL/0077), this sample size provided a 90% chance
`of showing a difference (at a 0.5% level of significance) in headache-
`response rate between the zolmitn'ptan S-mg nasal Spray dose compared to
`placebo 15 minutes after treatment. Calculations were based on the
`assumption that the headache response rate at 15 minutes would be 6% for
`placebo and 11% for zolmitriptan S-mg nasal spray.
`
`0 Sample Size for Interim Analysis: For this Interim Analysis, a correction
`for Type I error for both the interim and final full study was planned. The
`alpha spending function methodology based on Hwang, Shih, and deCani
`(1990) y-famiiy approach was used to control the overall two-sided type I
`error rate at 5% for both the interim and final analysis. The 2-sided
`significance boundaries for the p—values were calculated and pre-specified to
`
`be 0.0027 and 0.0479 (based on y=-2 and information fraction of t=l 5%) for
`the interim and the final analysis, respectively. That is, the statistical
`significance of the analysis results for the primary efficacy parameter of 2-
`hour headache response was tested against significance level of 0.0027 for
`the interim and 0.0479 for the final analysis.
`
`- The first 210 adult men and women who treated the first migraine attack
`with study medication, and who had an established diagnosis of migraine
`headache, with or without aura, as defined by 11-18 criteria were included in
`this analysis. This sample size provided approximately 90% power of
`showing a difference (at 0.27% two-sided level of significance) in headache
`response rate between the zolmitriptan S-mg nasal spray dose and placebo at
`2 hours after treatment. Calculations were based on the assumption that the
`headache response rate at 2 hours would be 39% for placebo and 69% for
`zohnitriptan S-mg nasal spray.
`
`3.1.1.6 Stratification
`
`There was no stratification for the full study.
`
`3.1.1.7 Interim Analysis
`
`The interim analysis results are presented in this statistical review.
`
`3.1.1.8 Efficacy Analysis Methods
`
`The sponsor described their statistical methods for the Interim Analysis as
`follows.
`
`

`

` STATISTICAL REVIEW AND EVALUATION
`
`The Interim Analysis of the efficacy endpoints was based on the principle of
`intention-to-treat (ITT). The ITT population was all patients who used trial
`treatment and provided baseline and post-baseline efficacy data for the same
`sweacy parameter, for first attack migraine attack treated. The primary efficacy
`endpoint for this Interim Analysis was headache response at 2 hours. Headache
`response is defined as an improvement in migraine headache intensity from
`severe or moderate to mild or none.
`
`The analysis plan for this Interim Analysis was based on the first attack data. For
`the primary endpoint of headache response at 2 hours, between-treatment group
`comparisons for the first attack were performed using the logistic regression
`method with treatment, region (mid-Atlantic, mid-West, New England, South,
`Southwest, West), and baseline intensity in the model. Due to the small sample
`size in the New England region, the statistical model did not fit properly;
`therefore, this region was merged with the nearest geographical region, the mid-
`Atlantic region, in the final model. The analysis results were presented in terms
`of odds ratios for the treatment effects, the associated 95% confidence intervals,
`
`and the corresponding p—values. The results of this statistical comparison were
`tested against a 2—sided significance level of 00027. No formal Interim Analysis
`was performed on headache response at 4 hours, 1 hour, 30 minutes, or 15
`minutes, but a summary of response rates at these time points was provided.
`
`Resolution of non-headache symptoms of migraine (nausea, photophobia, and
`phonophobia) were subjected to formal interim statistical analysis at 2 hours. The
`same logistic regression method with treatment, region, and baseline intensity in
`the model as for the analysis of the 2-hour response was used for these analyses.
`Since these were secondary analyses, nominal p—values were provided and no
`statistical adjustment for these secondary comparisons were made.
`
`3.1.1.9 Sponsor’s Results and Statistical Reviewer’s Findingleomments
`
`The sponsor defined the ITT population as all patients who used tn'al treatment
`and provided baseline and post-baseline efficacy data for the same efficacy
`parameter, for first attack migraine attack treated. Therefore, the sponsor’s
`efficacy claim was based on the ITT population which was not including two
`patients randomized but not evaluated at 2 hour: One patient had missing data at 2
`hour and the other patient had taken escape medication before 2 hours when
`migraine headache pain was mild. This statistical reviewer used LOCF algorithm
`for the efficacy analysis, which included the two patients.
`
`3.1.1.9.1 Baseline Characteristics
`
`f“.
`
`
`
`

`

`
`
` STATISTICAL REVIEW-AND EVALUATION
`
`Table 3 shows the key baseline demographic characteristics for the ITT
`population by treatment arm. All characteristics showed balance between the two
`treatment arms _
`
`A total of 210 patients with a history of migraine headaches were randomized into
`this study and included in the Interim Analysis. Of these, 108 patients received
`zolmitriptan and 102 received placebo. All 210 patients treated a migraine attack
`with at least 1 dose of study medication and provided at least 1 set of baseline and
`post-baseline assessments.
`‘
`
`Table 3.
`
`Demographic characteristic
`
`Demographic characteristic —— ITT populatitm
`'
`Treatment Eoup
`ZolmitriPWn 5.1113 nasal
`spray
`(§=1081
`
`A89 (3')
`Mean
`
`38.6
`
`Sex, number of patients (%)
`Women
`Men
`Race, number of patients (%)
`Caucasian
`_ Black
`Hispanic
`Asian
`Other‘
`Average number of attacks per month‘1
`3.7 (1.1)
`Mean (SD)
`2 — 6
`Range
`The sponsor’s results confirmed by the statistical reviewer.
`' Other includes any special subgroups.
`b This applied to the 3 months prior to study entry.
`
`89 (82.4)
`19 (17.6)
`
`72 (66.7)
`28 (25.9)
`4 (3.7)
`2 (1.9)
`2 (1.9)
`
`‘
`
`Placebo
`(lj=102!
`
`40.4
`
`87 (85.3)
`15 (14.7)
`
`69 (67.7)
`29 (28.4)
`4 (2.9)
`1 (1.0)
`0
`
`3.6 (1.1)
`2 - 6
`
`3.1.1.9.2 Primary Efficacy Analyses
`
`Table 4 gives the summary of efficacy results of primary endpoint, the 2-hour
`headache response of the first treated attack, for the FIT population. A total of
`108 patients on the zolmitriptan nasal Spray 5 mg arm and 102 patients on the
`placebo arm were included in the efficacy analysis. The number of patients with
`headache response at 2-hour were 76 (70.4%) and 48 (47.1%) in the two arms,
`respectively. There was a statistically significant difference (p=0.0006) between
`the two treatment groups.
`
`
`
`10
`
`

`

`
`
`
`
`
`1' STATISTICAL REVIEW AND EVALUATION
`
`Table 4.
`
`Primary endpoint: Headache response at 2 hours -—- first
`attack zolmitri tan nasal s ra vs.
`lacebo : FDA Anal sis
`
`.
`
`Zolmitriptan nasal spray 5.0 mg
`
`Placebo
`
`Patients evaluated at 2 h (n)
`Patients with headache response at 2 h (%)c
`
`(N=l08)‘
`108
`76 (70.4)
`
`(N=102l
`100
`48 (47.1)
`
`P-valuca
`Odds ratio:
`
`0.0006
`2.67
`
`{1.5, 4.9!
`95% confidence limitsa
`Statistical reviewer’s results based on the analysis data sets provided by the sponsor.
`' Patients received trial medication and treated a migraine attack.
`b Two patients from placebo group were not evaluated: One patient had missing data at 2 hour; the other
`patient had taken escape medication before 2 hours when migraine headache pain was mild.
`c Headache response is as a reduction in headache intensity from moderate or severe to mild or none.
`‘1 P-value is calculated by chi-square test.
`° Odds ratio and 95% confidence limits (C!) are estimated by SAS FREQ procedure.
`
`Reviewer '5' Comments:
`
`The sponsor’s efficacy claim was based on the ITT population which was not
`including two patients randomized but not evaluated at 2 hour: One patient had
`missing data at 2 hour and the other patient had taken escape medication before 2
`hours when migraine headache pain was mild. This statistical reviewer used
`LOCF algorithm for the efficacy analysis, which included the two patients.
`
`3.1.1.9.3 Secondary Efficacy Analyses
`
`Table 5 and Table 6 summarize the secondary efficacy analyses of the headache
`response at 15 minutes, 30 minutes, 1 hour, and 4 hours using the first attack, and
`the three migraine-associated symptoms: nausea, photophobia and phonophobia
`for ITT population.
`
`Table 5 shows that there was no significant difference for headache response
`between zolmitriptan nasal spray 5.0 mg and placebo arms at the 15 minutes and
`30 minutes (p-value=0.407 and 0.033, respectively) and there were significant
`differences between zolmitn'ptan nasal spray 5.0 mg and placebo at 1 hour and 4
`hours (p-value=0.019 and <.0001, respectively). The probability of the type I
`error had a decreasing trend when the observed time was increased.
`
`Table 6 shows that there were no significant differences for nausea, phot0phobia,
`and phonophobia symptoms between zolmitiiptan nasal spray 5.0 mg and placebo
`at 2 hours (p-value=0.080, 0.026, and 0.170, respectively).
`
`11
`
`

`

`STATISTICAL REVIEW AND EVALUATION
`
`Table 5.
`
`Secondary endpoint: Headache response at each time point
`zolmitri tan nasal s ra vs.
`lacebo : FDA Anal sis
`
`15 minutes
`
`,
`
`-
`
`Headache response at the time point ("/o)5
`p-value"
`Odds ratio (95% CI)‘1
`30 minutes
`
`Headache response at the time point ("/n)E
`1:3-valuec
`Odds ratio (95% Cl)‘1
`1 hour
`
`-
`
`Headache response at the time point (%)B
`p—value°
`Odds ratio (95% Cl)“
`4 hours
`
`Zolmitriptan nasal spray 5.0 mg
`
`Placebo
`
`(N=108)'
`
`(N= 102)
`
`18 (16.8)
`0.407
`1.33 (0.5, 3.3)
`
`40 (3 7.4)
`0.033
`1.92 g 1.0, 3.7)
`
`58 (54.7)
`0.019
`1.94 (1.1, 3.5)
`
`13 (12.8)
`NA
`NA
`
`24 (23.8)
`NA
`NA
`
`38 (38.4)
`NA
`NA
`
`52 (52.5)
`84 (79.3)
`Headache response at the time point (%)E
`NA
`< .0001
`p-value”
`
`Odds ratio (95% CI)d NA 3.45 (1.3, 6.7)
`
`Statistical reviewer’s results based on the analysis data sets provided by the sponsor.
`' Patients received trial medication and treated a migraine attack.
`b Headache response is as a reduction in headache intensity from moderate or severe to mild or none.
`c P-value is calculated by chi-square test.
`d Odds ratio and 95% confidence limits (CI) are estimated by SAS FREQ procedure.
`
`Table 6.
`
`Secondary endpoint: Nausea, photophobia and phonophobia at 2
`hours (zolmitriptan nasal vs. placebo):
`FDA Analysis
`Zolmitriptan nasal spray 5.0 mg
`Placebo
`
`Nausea
`
`(N=108)‘
`
`(N=102)
`
`28 (27.7)
`NA
`NA
`
`19 (18.6)
`0.080
`0.56 (0.3, 1.1)
`
`Patients with associated symptom at 2 h (%)E
`p-value‘
`Odds ratio (95% CI)‘1
`Photophobia
`50 (49.5)
`37 (34.3)
`Patients with associated symptom at 2 h (%)
`NA
`0.026
`p-valuec
`NA
`0.53 (0.3, 1.0)
`Odds ratio (95% CI)‘1
`Phonophobia —
`Patients with associated symptom at 2 h (%)
`29 (26.9)
`36 (35.6)
`p-value"
`0.170
`NA
`Odds ratio (95% CI)‘1
`0.66 (0.4, 1.2)
`NA
`Statistical reviewer‘s results based on the analysis data sets provided by the sponsor.
`' ' Patients received trial medication and treated a migraine attack.
`b Headache response is as a reduction in headache intensity from moderate or severe to mild or none.
`‘ P-value is calculated by chi-square test.
`d Odds ratio and 95% confidence limits (C1) are estimated by SAS FREQ procedure.
`
`12
`
`

`

` STATISTICAL REVIEW AND EVALUATION
`
`3.1.1.10 Sponsor’s Conclusions and Reviewer’s Conclusions/Comments
`
`Sponsor '5 Conclusion and Comments:
`
`The proportion of patients who had a response at 2 hours after the initial dose in
`the zolmitriptan treatment group significantly exceeded the proportion of who had