`
`Modafinil appears to have a relatively benign safety profile. Although there were occasional
`cases of clinically significant tachycardia and hypertension clearly related to study drug,
`in the
`placebo controlled trials the frequency of these events was similar in the modafinil and placebo
`treatment groups and most of the clinically significant events occurred at doses higher than the
`upper limit recommended by the sponsor in their proposed labeling. There was one case of
`psychosis undoubtedly related to study drug; but there was no evidence of an increased frequency
`of this AE in modafinil treated subjects at even the highest doses.
`
`Elevated GGT levels did appear to occur with increased frequency in the modafinil treated
`subjects compared to the placebo treated subjects. in a dose dependent manner. However, none
`of the elevations were markedly high, a number of the subjects had elevation at screening or
`baseline, some had other causes likely to explain the elevation (medical history, current
`medical problem, and/or concomitant medications), and a number of subjects showed decreasing
`values when they remained on drug.
`In addition, elevation of GGT in the absence of other liver
`enzyme elevations and/or elevation of bilirubin may well be due to liver induction. a
`documented effect of modafinil.
`
`in a single small
`One other issue related to safety should be addressed at this point.
`pharmacokinetic study in the elderly, MOD-O20, the Cm“, measured after a single dose of 300
`mg of modafinil, was almost twice as high as the Cmam measured after the same dose in healthy
`young subjects in Study MOD-018. There did appear to be a high incidence of AE’s in the
`nervous system early in the course of MOD-020, but the patient population in this small, open
`label trial consisted of hospitalized, elderly subjects with “behavioral problems”. making
`attribution difficult at best. The sponsor has included an upward dose titration for all patients
`in the recommended dosing schedule, noting in particular the need for that regimen in the
`elderly. The regimen should provide an adequate margin of safety in the elderly population.
`
`APPEARS THIS WAY ON ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`
`
`96
`
`'
`
`APPEARS THIS WAY 0N ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`/S/
`
`
`
`September 30, 1997
`
`gic Drug Products
`
`0C:
`
`NDA 20-717
`HFD-120 File
`H F D- 1 20
`Leber
`Katz
`
`RaPPaport
`Malandrucoo
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`
`
`APPENDIX l
`
`' ’
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`—
`
`

`

` .
`
`
` .
`
` .
`
`98
`
`FIGURE 1
`
`149 Studies Overall\
`/
`ln lntegrated Summary
`of Safety Database
`
`N = 95
`
`Not In lntegrated Summary
`of Safety Database
`
`Cephalon
`N814
`
`Non-Cephalon
`N881
`
`Cephalon
`~30
`
`Non-Cephalon
`54:54
`
`(
`
`._
`
`~
`
`A
`Clinical Pharmacology
`Narcolepsy
`Controlled
`Uncontrolled
`Non-Narcolepsy
`Controlled
`Uncontrolled
`
`-
`
`Completed
`N = 14
`
`Completed
`N = 81
`27
`5
`2
`3
`49
`31
`18
`
`Completed
`N = 45
`
`y
`
`‘
`
`N = 9
`
`Ongoing"
`o
`5
`2
`3
`4
`4
`0
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`" """
`
`""°'“°“
`
`
`
`

`

` .
`
` .
`
` .
`
`99
`
`Econcménaoo
`
`
`
`_.nausea—max
`
`EqQantassg
`93333.5
`
`Etugaggb
`«2.5525;er
`
`«ma—as
`
`Om?».2.0:358:35u233:8
`=O_uuuu\~e~\z>\Cm.3Daumnhg—xzzcx...
`
`Amnaflxuna~982.:_$.95
`OGSEBEQE
`EoquwanmoEn:SOEEE=>mEm
`
`
`
`Canaan«39.8
`
`
`
`«88.59.03”»?a35.3
`
` .
`
`
`
`BEST POSSIBLE COPY
`BEST POSSIBLE
`
`
`
`
`
`
`
`

`

` .
`
` .
`
`
`
` .
`
`
`
`
`“phalon-Sponsored Studies1tude;neluded in e
`of Safety Database)
`
`
`
`
`
`
`
`
`Clinical Pharrnacolgx Studies (12 stugigg
`
`C1538a/301lNA/US
`
`C1 538al3OZINAIUS
`
`
`
`
`
`
`
`
`
`
`
`
`cap-2101
`C1538a/108/NAIUK
`
`01538a/102/MT/us
`c153Ba/109lPK/UK
`
`C1538a/103IPK/US
`c1530a/11OIBE/UK
`
`C1538b/105/BE/UK
`c1saaa/111/PK/US
`
`C1538a/106/MD/US
`C15383l201IABIUS
`
`C15383/107IPK/UK
`c1ssaa/202/NA/us
` Controlled Narcolegx Studies (2 studies}
`
`
`
`
` ies (81 Studies; included in Integrated
`
`
` Cllnlgl Pharmafllggx Studies (27 gugigg)
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`5803
`MOD-O15
`M00001
`MOD-016
`1.100002
`M00017
`M00003
`M00018
`MOD-004
`' MOD-O19
`M00005
`MOD-020
`MOD-006
`M00021
`MOD-007
`M00022
`MOD-008
`MOD-023
`MOD-009
`MOD-029
`MOD-O10
`MOD-030
`M00011
`91424
`M00012
`P15951PK5
`
`A
`
`'
`
`'
`
`
`
`
`
`wwmmwc
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`1
`
`,
`
`

`

` .
`
`
`
` .
`
` .
`
`TABLE 1 (contd.)
`
`Iinical Pha
`
`a
`
`dies
`
`
`
`
`‘- Non-Cephalon-Sponsoredmm Case Report Forms (45 tis:
`not included in Integrated Summary of Safety database)
`‘
`
`
` DP-96-014 (A)
`
`
`88311-8
`Moo-006 (PCZ)
`690/1 -9
`
`
`MOD-013 (807IPCS)
`75311-10
`T1
`70711-11
`
`
`60111-2
`680/1-12
`
`
`OPEN/1-3
`70218-3 + 70818-3
`
`
`89011-4
`703/8-4
`OPEN/1-5
`74018-5
`
`
`689/1-6
`‘Tz
`
`
`694/1-7
`T3
`
`
`
`
`Uncontrolled Narcolegg Studies (1 study)
`
` Retrospective Compassionate Use
`
` Other Connolly Studies (17 studies)
`
`
`
`
`
`
`800/6-5
`705/5-5
`
`734/5-6
`733/53
`696/57
`7125-4
`69215-8
`85117-10
`79815-9
`687/7-7
`724/5-10
`71017.8
`
`685/6-1
`
`698/6-2
`
`797/6—3
`herU ntrolleds
`‘
`‘
`
`
`
`OPENfl-O
`OPENfl-B
`
`OPEN/54
`OPEN/$2
`
`
`
`Gama-111nm
`
`BEST POSSIBLE COPY ‘
`BEST POSSIBLE COPY
`
`,,
`
`

`

` .
`
` .
`
` .
`
`E1033
`
`cher antrolled Studieg (4 gggiesl
`
`E1029
`
`E1030
`
`E1032
`
`APPEARS THIS WAY 0N ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`Gamma; no:
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`

`

`104
`
`APPENDIX II
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`,’_______________—-_
`
`

`

`105
`
`
`
`wu<mw4n<hhm<4nwu<nm4n<h
`
`
`
`I'll-IIIIIlullvvIn1..III-uItIIaucnlIIIIIIIIIIUIIIIIIQnIIn!lllllIII-IIIcII-Io-nlao-Iluunllalvl-II
`
`
`
`
`
`
`
`annex.aeLozwou_s_4Loan:-=4:cocoa.nego:*ou_s_d5030‘uadd
`
`
`«wand
`
`
`
`5.:5.3%mung:
`
`”mouadu
`
`
`
`doxuzcomu.
`
`condo
`
`+m"A
`
`uwzohwx
`
`.n"A
`
`9.9Ao.ev3;
`
`x.»»o«g
`
`da\uzco.u.
`
`us.
`
`ma=\.nu.
`
`on:
`
`
`
`ugz\.n"A
`
`
`
`omo..A.oo..v>h_><¢au_._uu¢u
`
`o.opAu.~uvu.)
`
`«.m»d‘:_¢=
`
`
`
`.uoac.ucou.>oodoh¢xu=
`
`
`
`o.c.m~wda<h
`
`wk.
`
`s.~-o~<9:._c.‘ouox
`
`
`oo>o=nom<m.h_¢uumu
`
`
`
`
`.2:;3_!Eau
`
`BEST POSSIBLE COPY
`BESTPOS$BLECOPY
`
`ITEM 8 02002
`
`

`

`\
`
`106
`
`
`
`
`
`
`
`omcna.aELoz$0~_E_dgone:u2‘:onto:.nsgoz‘ou.e.dguaoduaddMm“
`
`
`
`
`
`
`
`
`
`
`2:AEv
`
`
`"A2:88:0 :58$55..I2a3:853:Wm.
`
`
`.n5xu:EnnT
`
`
`m..."y..53.85..S
`2cn.B
`Nu3L
`328.35..
`m.u
`E
`
`8e
`
`
`
`5:35.:5m5&3
`
`
`
`:z:
`
`
`.Eactcoe:3.sz:22:
`
`32u3:Xx2325av
`
`
`
`ESE.52
`
`8325.“:
`
`
`
`2:858“BaA2:38:
`knuvx583.2.E2uA
`P
`
`
`
`
`
`
`
`~uu<aWEE
`
`
`gmtank—«uwwbq<>£5253:szEIESM>3<u_2:uuz—z—xxwhwa
`c.c.m~WEE
`
`
`
`.2:£03.35‘
`
`$5.8<9.:ctuua:
`
`
`
`3528292513
`
`
`
`BEST POSSIBLE COPY
`TI
`L
`0
`C
`S
`
`10..0
`
`
`

`

`107
`
`2.5cn"A
`
`2.5cn"A
`
`2.5cn"A
`
`mm"A
`
`on"A
`
`m.:A
`
`m:A
`
`awn"A
`
`°.~uA
`
`ascnuA
`
`00..A
`
`ONuv
`
`«.0v
`
`00v
`
`
`
`35.538:ms.33.:
`
`2:59;
`
`:83:<
`
`
`
`2193hm<
`
`m»$62.3_a
`
`x335
`
`we8...:-
`
`23¢
`
`35830:“.
`
`ma—2:55
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`ITEM 8 020m
`
`
`
`
`
`aha—:23.3022.6
`
`
`
`Fuu<awdn<~
`
`
`u.~.o~«a:__c_*muo:
`oa>ozaom<m.h~«uumu
`
`
`.2:63238
`
`
`
`

`

`Memorandum
`
`Department of Health and Human Servrces
`Public Health Service
`Food and Drug Administration
`Center for Drug Evaluation and Research
`SEE“““E£71.32?‘?€“1§37‘““““‘“““““-\‘*“
`FROM
`Paul Leber, MD
`Director,
`
`HFD-120
`
`Division of Néuropharmacological Drug Products
`Provlgil Tablets
`(modafinll)
`
`NDA 20-717
`
`SUBJECT
`
`TO:
`
`—_____—-__
`
`F
`
`Excessive daytime sleepiness, episodes of “sudden and Irre3lstible" sleep
`onset m inappropriate settings drop attacks due to the sudden loss of
`
`muscle tone (aka 0
`
`'
`
`BEsT POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`
`
`— ‘
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`

`

`page 3 of 13
`Leber: Provigil®[modafinil tablets] Approvable Memo
`wakefulness;
`the decrement in GABA is associated with increases in the
`local release of both dopamine and glutamate.
`it has been suggested that
`an increase in the concentration of either of the latter neurotransmitters
`
`importance to
`different profiles of pharmacologic effects is of potential
`the former’s commerciahpromotion;
`it
`is not so clear whether these
`differences predict any clinical advantage, however.
`
`Administrative Review Issues
`
`is the subject of a pending FDA Recommendation that it be
`Modafinil
`placed in Schedule lV under the CSA.
`
`Statistical Review
`
`David Hoberman, Ph.D.
`
`Aisar Atrkchi, Ph.D.
`
`Joseph DeGeorge, Ph.D.
`
`
` Team Leader memorandum Russell Katz M.D.
`
`12/4/97
`Supervisory Pharmacology Glenna Fitzgerald, Ph.D.
`11/7/97
`
`
`
`Memorandum
`
`
`
`
`i
`
`
`
`
`10/31/97 ‘
`m Wendelyn Schmidt
`
`
`1 1/19/97
`Biopharm
`Rae Yuan, Ph.D.
`
`
`12/2/97
`)_ Martha Heiman, Ph.D.
`
`
`
`
`
`Draft CSA Scheduling
`
`
`12/1/97 J
`Michael Klein, Ph.D.
`Recommendation
`
`‘K _M-‘a,_1\\___
`
`
`
`
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`-‘ .._.__. wv....‘.-__-" .s -a
`
`page 4 of 13
`
`y. Modafinil has not been shown to
`
`’
`
`Study Design
`
`The two studies relied upon wer
`as Study 301 and Study 302.
`
`e conducted in the US; they are identified
`
`the subject at each visit.
`
`The primary statistical analysis for the M
`covariance employing terms for study site and baseline MWT.
`
`No imbalance in groups was foun
`there were minor differences in the drop-
`extent of differential attrition was at no
`point considered an important
`factor in the interpretation of either trial.
`The actual retention of
`subjects by time and by treatment is provid
`ed in Table 1
`taken from the
`appendix to Dr. Hoberman’s review that fol
`lows below.
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`

`

`page 5 of 13
`
`
`
`APPEARS THIS WAY ON ORIGINAL
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY—
`
`

`

`page 6 of 13
`
`ON
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`E a 3 E
`
`"
`'3
`
`Leber: Provigil®[modafinil tablets] Approvable Memo
`
`
`
`Study CISMSBDZINAIUS
`
`STUDY
`
`302
`
`

`

`
`
`Leber: Provigil®[modafinil tablets] Approvable Memo
`
`page 7 of 13
`
`The CGl-C results provide further support for a beneficial effect of
`
`modafinil.
`
`
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`—
`
`

`

`( about the global notwithstanding,
`
`reasonable and responsible to
`
`i am persuaded that it is both
`' d. Accordingly, my uncertainties
`'
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`
`
`Leber: Provigil®[modafinil tablets] Approvable Memo
`
`page 9 of 13
`
`support a regulatory conclusion that a drug is safe for use.
`
`Tumorigenicity testing is
`
`incomplete
`
`A second in-vivo life
`inducing potential.).
`-time study was conducted in
`the mouse, but at system'ic exposures deem
`ed inadequate by both the
`Division review team and the CAC.
`There is no simple fix for this
`deficiency, however.
`
`induces its’own metabolism in the mouse to an extent
`First, modafinil
`that cannot be overcome by merely increasing the administered dose.
`Accordingly, another preclinical
`tumor assay is required.
`However,
`because modafinil
`is non-genotoxic, alternatives are limited.
`In fact,
`according to Dr. Fitzgerald,
`that
`the only viable alternative is the TG.AC
`assay because it is the “only mbdel for non-
`'
`
`time CA and TG.AC assays are fun ible at l
` .
`
` .
`
`status is concerned
`
` .
`
`Marginal Teratogenicity and Reproductive performance
`testing
`
`The rat fertility study and
`ere not done under GLP.
`these studies were conducted at sub-
`
`
`
`3 page 2, Dr. Fitzgerald’s November 7, 1997 Supervisory Overview
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`

`

`Leber: Provigil®[modafinil tablets] Approvable Memo
`
`page 10 of 13
`
`Furthermore, final reports of the GLP compliant peri and
`optimal doses.
`post-natal study, a study ordinarily required for NDA approval, has yet to
`be submitted.
`
`These deficiencies notwithstanding, Dr. Fitzgerald concludes that she can
`recommend an approvable action because 1) the rat teratogenicity study,
`albeit conducted at inadequate doses, did establish a threshold dose for
`teratogenic effects (based on body surface area,
`the threshold teratogenic
`dose is only 5 fold that recommended for human use) and 2) preliminary
`reports of the GLP compliant study indicate that modafinil
`is unlikely to
`have adverse behavioral effects on newborns.
`
`In sum, although modafinil's teratogenic effects and its effects on
`reproductive performance have not yet been fully evaluated, Dr. Fitzgerald
`believes that
`these limitations in its assessment are not sufficient,
`in
`and of themselves,
`to justify a not approvable action.
`Accordingly, Dr.
`Fitzgerald recommends an approvable action with final approval
`contingent upon the sponsor’s agreement to conduct and submit reports in
`phase IV of 1) a TG.AC assay and 2) GLP compliant teratogenicity and
`reproductive performance studies.
`It
`is also evident that Dr. Fitzgerald
`expects that product
`labeling will
`identify the deficiencies and
`limitations of the tests conducted to date.
`
`Clinical Safety
`
`typically reflects
`A regulatory determination that a drug is ‘safe for use'
`a consensual judgment, offered by a team of agency physicians and
`scientists,
`that the risks reported in association with the use of a new
`drug are,
`taking into account their kind, severity and incidence, reasonably
`acceptable in light of the benefits likely to accrue from the use of the
`drug under the conditions recommended in its proposed product labeling
`
`Although NDA approval always implies a warranty that the product that is
`the subject of that NDA will be ‘safe for use,’ under the conditions of use
`proposed in product labeling, the strength (and value) of the warrant is.
`very much affected by the quality and quantity of clinical experience
`gained with the drug product during its premarketing development.
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`page 11 of 13
`
`BEST POSSIBLE COPY
`—
`BEST POSSIBLE COPY
`
`

`

`
`
`/S/
`
`BEST POSSIBLE COPY
`— BEST POSSIBLE COPY
`
`
`
`

`

`Leber: Provigil®[modafinil tablets] Approvable Memo
`cc: NDA 20-717
`
`Page 13 of 13
`
`HFD-101
`
`Temple
`HFD-120
`
`Katz
`27532533
`
`Atrakchi
`‘ Guzewska
`Heimann
`HFD-71O
`
`Hoberman
`
`Salhroot
`HFD-860
`
`Baweja
`HFD-17O
`
`Klein
`McCormick
`
`,
`
`~
`
`APPEARS THIS WAY ON ORIGINAL
`
`‘
`
`»
`
`~
`
`( f'.
`3.
`
`\ -
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`—
`
`

`

`BACKGROUND
`
`for the use of Modaflml m
`ment of Patients With Narcolepsy
`
`NDA 20-717 was submitted by Cephalon. Inc. on 12/30/96
`
`Modafinil as a
`
` .
`
` .
`
` .
`
` .
`
` .
`
` .
`
` .
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`
`
`Q ~
`
`sponsored, and submitted the results of, 2
`'
`'
`controlled trials which the
`These studies were of essenti
`effectiveness.
`STUDY 301
`
`BEST POSSIBLE COPY
`—
`BEST POSSIBLE COPY
`
`

`

`
`by the American Sleep Disorders Association (published in 1990) were
`eligible for entrance into the study.
`Specifically, patients had to have
`met one of the following 2 minimal criteria (from Dr. Rappaport’s review,
`
`(M
`
`for at least 3 months, plus sudden bilateral
`
`loss of postural muscle tone
`
`B. Complaint of sudden muscle weakness or excessive sleepiness, plus
`sleep paralysis, hypnagogic hallucinations, automatic behaviors, and
`disrupted major sleep episodes; plus polysomnography’ showing either: 1)
`sleep latency less than 10 minutes. or 2) REM latency less than 20
`minutes, and 3) an MSLT demonstrating a sleep latency of equal to or less
`than 5 minutes, and 4) 2 or more sleep onset REM periods; plus no other
`medical or psychiatric disorder that could explain the clinical symptoms.
`
`1) Epworth Sleepiness Scale-a scale designed to assess sleepiness in
`8 different daytime situations. The scale ranges from 0 (normal) to 24
`(worst daytime sleepiness).
`
`(SCPT) training session and one
`2) “Steer Clear" Performance Test
`SCPT-a computerized test that measures the subject’s ability to avoid
`
`followed by a Multiple Sleep
`3) Two polysomnography sessions; 1
`Latency Test
`(MSLT), and 1 followed by a Maintenance of
`Wakefulness Test
`(MWT):
`
`a) MSLT-in this test,
`
`the patient is instructed to lie quietly and attempt
`
`3
`
`-
`
`-
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`g
`
`

`

`to fall sleep. Sleep latenc
`y is the time to the first 16 seconds of any
`stage sleep.
`in addition,
`alternate criteria for the determination of sleep
`latency were used;
`this wa
`inued sleep latency.
`Other sleep parameters are
`also measured. Four of
`these tests are
`performed in each session.
`Each test is terminated
`at 20 minutes if no
`sleep has occurred.
`
`,
`the
`b) MWT-ln this test,
`in bed in a dark room and '
`measured in 4 separate tests. Each test
`not occurred in 20 minutes.
`
`'recumbent on a reading pillow
`'
`will be terminated if sleep has
`
`4) Patient’s Dally Sleep Log-Patients recorded multiple sleep related
`measures; e.g.,
`total minutes of daytime sleep, number of night time
`xy. hypnogogic hallucinations,
`
`5) Quality of Life in Nar
`designed instrument to as
`functioning.
`it
`is divided
`
`Inventory-A Cephalon
`colepsy (QOLIN)
`sess the effect of narcolepsy on the patient’s
`into 5 components. each with multiple
`on an ordinal or visual analogue scale.
`Impression of Severity (CGl-S)
`6) Clinical Global
`After baseline, patients were randomized to receive one of 3 treatments,
`each given as a single daily dose (4 tablets):
`
`Modafinil 200 mg/day
`Modafinil 400 mg/day
`Placebo
`
`There were 2 primary outcome measures:
`
`1) The Maintenance of Wakefulness Test (MWI’), which made 4
`determinations of sleep latency at each visit Baseline, Weeks 3, 6. and 9).
`2) Clinical Global Impression of Change (CGl-C) - a 7 point symmetric
`scale, centered around no change, and ranging from very much improved to
`
`4
`
`.
`
`-
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`—
`
`

`

`k
`
`A. _.~_ .N--.~___n7-.M_' n... ._ _
`
`.e
`
`4
`
`»
`
`very much worse.
`
`sided at the 5% level. The population to be included in the analysis was to
`be all patients who receive study medication who had at least one post-
`baseline assessment that consisted of both of the primary measures.
`The primary analysis was to be of the last assessment for each patient.
`The primary method of analysis of the sleep latency (MWT) was to be a
`generalized ANCOVA which was to include treatment group, center,
`baseline sleep latency, and other covariates chosen to be determined by a
`stepwise selection procedure. This procedure was designed to identify
`any covariates that show baseline variability that significantly ..correlate
`with sleep latency at endpoint and any covariates that show clinically
`significant variability at baseline (description taken from Dr. Rappaport's
`review, page 19).
`
`(f ..
`‘
`
`The primary analysis of the CGl-C was to be a logistic regression model
`with terms for treatment, center, baseline severity, and other covariates
`to be selected as described above.
`
`Other aspects of the data collected on the previously described
`instruments were to be analyzed as secondary outcomes.
`
`APPEARS THIS WAY 0N ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`—
`
`

`

`Placebo
`Modafinil 200
`Modafinii 400
`94
`96
`95
`Patients randomized
`92
`96
`95
`Patients treated
`Patients in Efficacy
`92
`95
`86
`Analysis
`87 (93%)
`93 (97%)
`81 (85%)
`Completers
`Patients were not Included in the efficacy anainIs as noted above
`because they did not have at
`
`APPEARS THIS WAY 0N ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`6
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

` _
`
`-_u _. “-.‘_a:..._._--.-.h-4 -- .-
`
`Primary Outcomes
`
`Sleep Latency (MWT)
`
`The following table displays the results:
`
`Baseline
`
`Endpoint
`
`P-value
`
`Placebo (N=92)
`M 200 mg/d (N=95)
`M 400 mg/d (N=86)
`
`l
`5.8
`5.8
`6.6
`
`5.1
`8.2
`8.9
`
`<0.001
`<0.001
`
`The analysis used was a simple ANCOVA with baseline latency as the only
`covariate (as described in the protocol, but with no additional covariates
`included).
`
`As can be seen in Dr. Hoberman’s Figure 1, these differences were
`apparent at Week 3 (the first on-treatment visit) and persisted with little
`change over the duration of the study. Dr. Rappaport's Table 14, page 32
`describes the same information, and notes that the p-values for each dose
`comparison to placebo at each of these timepoints was <0.001.
`
`APPEARS THIS WAY ON ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`m
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`

`

`CGl-C
`
`There was a marked change in the distribution of change scores, as
`outlined in the following table:
`
`Modafinil 400 Modafinil 200
`(N=86)
`(N=95)
`
`Placebo
`(N=92)
`
`Very much improved
`Much Improved
`Minimally Improved
`No change
`Minimally Worse
`Much Worse
`Very Much Worse
`
`t
`
`9%
`41%
`22%
`23%
`3%
`1%
`0%
`
`7%
`26%
`31%
`28%
`5%
`2%
`0%
`
`4%
`9%
`24%
`47%
`12%
`3%
`1%
`
`P-value
`
`<0.001
`
`<0.001
`
`The model used baseline severity as a covariate with no other covariates
`used.
`
`Although the documents submitted with the NDA were not explicit
`regarding the independence of the rater of the CGI,
`I discussed this in
`great detail with the sponsor in a phone call on November 19, 1997. They
`assured me that the rater (usually, but not always, a physician) had no
`access to any post-baseline data about the patient at the time of the
`ratings.
`Further, every effort was made to ensure that the rater remained
`the same throughout the study.
`
`As can be seen in Dr. Hoberman’s Figure 2, these differences were seen by
`Week 3, and remained essentially unchanged throughout the trial. Also,
`the proportion of patients who improved at endpoint was significant at
`p<0.001 for each dose group compared to placebo (72%, 64%. 37%, for
`Modafinil 400, Modafinil 200. and Placebo. respectively).
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`st 10 secon s (a standard'measure), REM sleep latency,
`total
`.
`ncy to sleep
`sleep time, and patients subjectlve evaluation of
`
`Agency has
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`

`

`
`
`APPEARS THIS WAY ON ORIGINAL
`APPEARS THIS WAY ON ORIGINAL
`
`10
`
`—
`
`BEST POSSIBLE COPY
`BEST POSSIBLE COPY
`
`
`
`