Page 69 of 89
`Ranjit B. Mani, MD, HFDo120 Medical Review
`
`NBA 21196, Xyrem , Orphan Medical, Inc. 6/15/01
`
`10.12 Analysis Plan
`
`1012.1 Demographic And Baseline Variables
`
`The 2 double- blind period treatment groups were to be compared in regard to
`demographic and baseline variables
`Quantitative variables were to be analyzed using either a ttest or a Wilcoxon
`rank sum test as appropriate
`Qualitative variables were to be analyzed using Fishers exact test
`
`10.122 Primary Efficacy Parameter
`
`The primary efficacy parameter was the change in the number of cataplexy
`attacks per week in the 2eweek period following Visit 3 (endpoint), compared
`with the 2-week period prior to Visit 3 (baseline). if a subject withdrew prior to
`Visit 4 the weekly average would be calculated based upon the data that
`were available
`
`The efficacy population was to consist of all those randomized at Visit 2 who
`had some post-baseline efficacy data
`The above change in the weekly number of cataplexy attacks was to be-
`analyzed using a non-parametric ANCOVA as follows
`.
`The baseline number of cataplexy attacks and the change in the weekly number of
`cataplexy attacks were to be replaced by their corresponding ranks (mean ranks will be
`used when ties occur).
`The ANCOVA would be constructed from the residuals derived from the ordinary least
`squares prediction of the change in the weekly number of cataplexy attacks based on a
`simple linear model
`The treatment groups would then be compared with respect to these residuals using the
`Wilcoxon rank sum test.
`~
`
`-
`
`.
`
`.
`
`Prior to completion of the analysis a test would be performed to compare the slopes for
`the 2 treatment groups.
`The significance of the mean change from baseline for each treatment group
`would be determined using the Wilcoxon signed rank test
`
`1012 3 Safety Parameters
`
`The safety population would consist of all those randomized to receive drug
`at Visit 3 who had some post--baseline safety data
`-
`Adverse events would be summarized by treatment group and organized by
`' preferred term and body system. Treatment groups would be compared to
`the incidence of each adverse event using Fisher’s exact test
`Laboratory data would be summarized in tabular form as well as with the use
`of shift tables. Treatment groups would be compared in regard to the mean
`change from baseline using ANOVA. Within each treatment group the
`significance of the mean change from baseline will be analyzed using a
`paired t-test
`
`10.1.2.4 Sample Size Rationale
`
`The sample size calculation was based on the change in weekly cataplexy
`attacks comparing the 2 weeks prior to randomization and the 2 weeks after
`randomization
`
`

`

`Flanjit B. Mani, MD, HFD-12O Medical Review
`NBA 21196, Xyrem , Orphan Medical, inc.
`
`Page 70 0t 89
`6/15/01
`
`0 The assumptions for the sample size calculation were as follows
`. Power of 80 %
`
`. 2-sided 0c of 0.05
`
`. A 50 % increase in the total number of cataplexy attacks in the placebo
`group, and a 10 % increase in a Xyrem® group
`. A standard deviation, based on a log transformation, of about 0.30 for
`the change in total number of cataplexy attacks (based on a previous
`study)
`0 Based on the above, a sample size of 22 patients would be required per
`treatment group to detect a treatment difference.
`. To allow for a minor departure from the above assumptions a total of 30
`patients would be randomized to each treatment group
`
`10.13 Protocol Amendments
`
`These have been incorporated into the above
`
`10. 14 Actual Analyses Performed
`
`10.15 Efficacy Results
`
`The study was conducted at 14 centers. Each center enrolled between 1 and 7
`patients
`
`10. 15. 1 Patient Disposition
`
`Patient disposition is summarized in the following schematic copied from the
`submission
`‘
`'
`
`
` SE
`
`PETEENTS 3C ESEEDIRASCCHTESB
`
`
`
`N=Ci
`N~0
`
`
`CQMPLETED
`WI?E3§AW§
`-?h -» ;
`WITESRRHH
`
`
`
`
`Note that 1 randomized patient failed screening because of concomitant use of a
`selective serotonin re-uptake inhibitor (paroxetine). The blind was broken on 1
`patient shortly after completion of the trial on account of a serious adverse event.
`
`

`

`Raniit B. Mani, MD, HFD-120 Medical Review
`NBA 21196, Xyrem , Orphan Medical, Inc.
`
`10.152 Protocol Deviations
`
`Page 71 of 89
`6/15/01
`
`. One patient was allowed into the trial despite having been treated with GHB
`for 3.7 years (the inclusion criteria specified that the duration of treatment
`should be from 0.5 to 3.5 years)
`. One patient was allowed to continue in the trial despite receiving bupropion
`as a medication for cataplexy
`3 patients overmedicated
`For “efficiency” 2 patients who were taking 3 g/day at study entry and
`continued to take that dose during the study were listed as taking 4.5 g/day
`For a number of patients Visits 1 and 2 were combined.
`
`o
`.
`
`.
`
`10. 15.3 Medication Compliance
`
`As the following table indicates medication compliance was comparable for the 2
`Phase III treatment groups
`
`Tris}. Xcdicazicn
`Adziniatzuticn
`
`lyre. (H326)
`Phat H mm— Phase II‘
`
`Placebo {11:29)
`
`
`
`P'IWCAKMUMEN'N
`
`10.154 Baseline And Other Demographic Characteristics
`
`These characteristics are summarized in the next 2 tables copied from this-
`submission. Although gender, and baseline frequency of cataplexy attacks were
`not entirely balanced between the treatment groups the sponsor describes-the
`differences as not being statistically significant. Note that the daily dose of
`Xyrem® did appear balanced between the Phase III treatment groups.
`
`APPEARS THIS WAY
`0N ORIGlNAL
`
`

`

`Ranjit B. Mani, MD, HFD-12O Medical Review
`NBA 21196, Xyrem , Orphan Medical, Inc.
`
`Page 72 of 89
`6/15/01
`
`lIhSS)
`
`xyru ”€426!
`
`Flat-b0 [26-29)
`
`p—Valu.
`
`Chataczzrmucs
`1:._ .‘",’~‘.}
`
`_\
`
`. ass
`
`’5
`
`. 2’?!
`
`.256
`
`Z::>n:i:ue§<fii
`
`6
`v iLcltfidlti
`
`mm amp
`Charactntilt£cx m Xyran was)
`Plan-bu m-zs)
`
`p-vuc-
`
`3:.
`
`10.15.5 Primary Efficacy Analysis
`
`An intent-to-treat analysis was performed as specified in the protocol comprising
`all patients who received one or more doses of trial medication during the double
`blind withdrawal period and had recorded baseline and post-baseline efficacy
`measures
`-
`
`The results of the primary efficacy analysis are outlined in the table and figure
`below. For those receiving Xyrem® during the double-blind withdrawal phase
`there was no median change from baseline in the number of cataplexy attacks
`over the 2 week period of withdrawal. For those receiving placebo during the
`withdrawal phase the median change in the number of cataplexy attacks during
`as compared with baseline showed an increase. The difference was statistically
`significant (p < 0.001). Note that the table and figure below depict median
`change
`
`

`

`Ranjit B. Mani, MD, HFD-120 Medical Review
`NBA 21196. Xyrem , Orphan Medical, inc.
`
`Page 73 of 89
`6/15/01
`
`xyrcn (H-Zéi
`
`Placeba (SuZS‘
`en
`
`25
`
`20
`
`_¢.50U"
`
`
`(Median!2Weeks)
`
`ChangeinCatapiexyAttacks
`
`pc 0001'
`
`Xyrem
`
`m.
`Piacebe
`
`
`
`rt»,:MCGVA msfitl castaznisg rank baseline, zrsdtman g:cup, nan
`' trcatzcn: grcup inicraczicx.
`
`As the next tabie and figure indicate the median change from baseline by week
`in the number of catapiexy attacks mirrors that for the primary efficacy analysis
`above
`
`m pun-m
`
`mum number a! 9&L4euts
`
`M23: 1 33
`age an
`
`Number :1 92:: ienis
`flex“ 1 83
`Median
`Xinimum
`Maximum
`End; in;
`ZS tutu-:4. :0 l
`
`m:
`
`.X;
`
`rm.
`
`,—..e...._.....
`
`-3, "awn“...m, 5..” ~Jx‘h “mu-w. Va -‘pupaulf AS..I~I...I
`
`r.
`
`.
`
`..
`
`s u A u”). humus
`
`

`

`Ranjit B. Mani, MD, HFD—120 Medical Review
`NDA 21196, Xyrem , Orphan Medical, Inc.
`
`Page 74 of 89
`6/15/01
`
`14
`
`_. N
`
`—L 0
`
`(Median!Week)
`
`
`
`
`
`ChangeinCalaplexyAttacks
`
`
`Week 1
`Week 2
`Double~Biind Period
`
`No formal analyses were carried out to evaluate differential effects at study sites,
`or to evaluate drug-drug or drug-disease interactions.
`'
`
`10.15.6 Analysis Of Secondary Efficacy Measures
`
`This study had no secondary efficacy measures
`
`10.16 Safety Results
`
`These are summarized in the NDA Safety Review.
`
`10.17 Sponsor’s Conclusions Regarding Efficacy
`
`Xyrem® is effective as a long-term treatment for cataplexy
`
`10.18 Reviewer’s Comments
`
`The design and analysis plan for this study were discussed at length a priori
`with this Division
`
`The Division had agreed earlier that the randomized withdrawal paradigm
`used in this study would be appropriate for demonstrating the long-term
`efficacy of Xyrem®. Based on that agreement and the study results, I would
`agree with the sponsor’s conclusion that this study provides evidence for the
`long-term efficacy of Xyrem®Jn the treatment of cataplexy.
`Dr Sharon Yan, Agency statistical reviewer, has informed me that she also
`agrees with the sponsor’s conclusion that the study provides evidence for the
`long-term efficacy of GHB in treating cataplexy.
`An effective daily dose of Xyrem® is difficult to determine from this study
`since patients were not randomized to separate Xyrem® dose groups prior to
`withdrawal. However, it is noteworthy that although the dose of Xyrem® used
`in this study ranged from 3 to 9 g/day, 80% of those enrolled were receiving
`doses of 6-9 g/day at study entry.
`
`At a Peripheral and Central Nervous System Advisory Committee meeting held on
`6/8/01 to discuss this NDA the sponsor did present a further analysis not included in
`
`

`

`Ranjit B. Mani, MD, HFD-120 Medical Review
`NDA 21196, Xyrem , Orphan Medical, Inc.
`
`Page 75 of 89
`6/15/01
`
`the formal submission. In this analysis those patients who received placebo during
`the randomized withdrawal phase were divided into subgroups based on the dose of
`Xyrem® that they were taking prior to withdrawal. Robust statistically significant
`differences were seen on these post-hoc comparisons between the placebo group
`subsets who were originally taking doses of 6 g/day, 7.5 g/day and 9 g/day, and
`those continuing to take Xyrem® during the randomized withdrawal phase. These
`results were interpreted as showing further evidence for the efficacy of doses in the
`6 to 9 g/day range in the treatment of cataplexy. Note that these subgroups were not
`randomized.
`
`11.» Study OMC-SXB-16
`The report of this study was submitted 12/16/00
`
`This study was carried out to compare the taste of 3 different placebo
`formulations with that of Xyrem® oral solution.
`
`The study was carried out on 12 healthy volunteers, all employees of Orphan
`Medical, Inc.
`
`Appropriate inclusion and exclusion criteria were used.
`
`Blind comparisons were made between sodium oxybate oral solution and the 3
`different placebo formulations which contained sodium phosphate, sodium
`chloride and sodium citrate, respectively. The placebo solutions to be tested
`were presented in random order. Comparisons were made in pairs: each of_the 3
`placebo solutions matched with the solution of Xyrem®. Each subject was '
`exposed to 6 different solutions
`
`The similarity of taste for each solution pair was evaluated using the Formulation
`Taste Test Questionnaire. The questionnaire consisted of a visual analog scale:
`a line exactly 10 cm in length with the extremities labeled “identical” and “very
`dissimilar.” Subjects were asked to complete the questionnaire, after swishing
`and expelling each solution of the blinded pair, by marking the visual analog
`scale.
`
`The Xyrem® dose administered was 3 g.
`
`The study concluded that the placebo formulated with sodium phosphate
`was the one most comparable to Xyrem® oral solution. The sodium -
`phosphate and sodium citrate solutions were considered acceptable
`candidates for placebo.
`
`No adverse events were observed during the study which was conducted in April
`1999.
`'
`
`Note that
`
`

`

`Ranjit B. Mani, MD, HFD-120 Medical Review
`NDA 21196, Xyrem , Orphan Medical, Inc.
`
`Page 76 of 89
`6/15/01
`
`For the OMC-GHB-2 efficacy study a sodium chloride placebo formulation
`was used
`
`For the OMC—SXB-21 efficacy study a sodium citrate placebo formulation was
`used
`
`12. Overall Comments Regarding Efficacy
`
`The sponsor is seeking a claim for Xyrem® as a treatment for cataplexy and
`daytime sleepiness accompanying narcolepsy
`The evidence for the efficacy of GHB in treating cataplexy may be
`summarized as follows
`
`There does appear to be evidence that GHB is effective in treating cataplexy,
`although there is currently no evidence that the drug is effective in treating
`complete cataplexy attacks, the most serious form.
`The evidence for efficacy is mainly established by the results of Studies OMC—
`GHB-2 and OMC-SXB-Zl, and to a lesser extent by the Scrima study which has
`a number of deficiencies. In all 3 studies the same outcome measure, the
`
`frequency of cataplexy attacks based on patient diaries, was used. The
`Lammers study must be considered a “negative” one at this time.
`The effective dose of GHB in treating cataplexy can be best defined from the
`OMC-GHB-Z study in which patients were randomized to specific doses of GHB
`and robust evidence of efficacy was seen only at a dose of 9 g/day (and not at 3
`g/day and 6 g/day). In the OMC-SXB-Zl and Scrima studies, there was no
`randomization by GHB dose: in the OMC-SXB~21 study 80% of patients had
`been taking GHB doses 2 6 g/day prior to randomized withdrawal. In the less
`than optimal Scrima study the protocol-specified dose was 50 mg/kg/day, and
`the mean daily dose estimated using body weight data was 4.5 g/day. Thus the
`most clearly effective dose in treating cataplexy was 9 g/day with less clear and
`consistent evidence of efficacy at lower doses 2 4.5 g/day. The evidence in favor
`of efficacy at the 6 g/day dose comes mainly from the OMC-GHB-Z study and is
`marginal and analysis-dependent. There is no evidence for efficacy at a dose of
`3 g/day
`The evidence for efficacy/or the lack thereof of GHB in treating daytime
`sleepiness accompanying narcolepsy may be summarized as follows
`Efficacy is supported by the sponsor’s analysis of the Epworth Sleepiness Scale
`in the OMC-GHB-2 study (although not by the analysis performed by Dr Sharon
`Yan, FDA statistician on the same measure in the same study), and to a very
`small extent by the analysis of the frequency of daytime sleep attacks in the
`Lammers study which had a number of inadequacies
`However the analysis of a number of other measures of daytime sleepinessIn 3
`efficacy studies could not be considered to show a statistically significant
`‘
`superiority of GHB over placebo. These included the following
`.
`The frequency of daytime sleep attacks and the duration of such attacks'In the OMC-
`GHB-2 study
`‘
`The Sleepiness Index (of the Multiple Sleep Latency Test), which was a primary
`efficacy measure, and the frequency of daytime sleep attacks, in the Scrima study
`The severity of daytime sleep attacks in the Lammers study
`'
`I
`
`.
`
`.
`
`

`

`Ranjit B. Mani, MD, HFD-120 Medical Review
`NDA 21196, Xyrem , Orphan Medical, lnc.
`
`Page 77 of 89
`6/15/01
`
`.
`
`o
`
`In the OMC-GHB-2 study the only seemingly effective dose in treating daytime
`sleepiness, based on the sponsor’s analysis, was 9 g/day. in the Lammers study
`the mean daily dose used was 4.75 g/day
`It is unclear to what extent the analysis of the Epworth Sleepiness Scale data in
`OMC~GHB~2 was confounded by the concurrent use of stimulant medication (it is
`unclear to what extent the treatment groups were matched in this regard; see
`Section 6.14.3.6)
`. The lack of replication of the effect of GHB on daytime sleepiness as assessed
`by a specific measure in more than one study is unsatisfactory, quite apart from
`the other deficiencies noted in the efficacy studies. As noted above, the most
`robust evidence in favor of
`,
`. Currently, there does NOT appear to be adequate evidence that GHB is effective
`in treating daytime sleepiness accompanying narcolepsy.
`Also note that this Center’s Division Of Scientific investigations carried out an
`inspection of the Scrima study site (see Section 15) and concluded that the
`data were unacceptable because most of the drug accountability records
`were missing. DSI recommended that data from the Scrima study not be
`used in support of the NDA
`in summary,
`. Evidence has been provided in this application that Xyrem® is effective in
`treating cataplexy. The evidence is best at a Xyrem® dose of 9 g/day, but may
`extend across the dose range of 6-9 g/day
`. The application does not however provide adequate evidence that Xyrem®IS
`effective at treating daytime sleepiness accompanying narcolepsy
`
`13. Labeling Review
`
`Please see separate document entitled “NDA 21196 Labeling Review”
`
`14. Overall Comments Regarding Safety Of Xyrem®
`
`See NDA Safety Review for full details
`
`14. 1 Clinical Safety
`
`When GHB is used to treat narcolepsy in doses of 3-9 g/day the most
`common, dose-related, and seemingly drug-related, adverse events haVe
`included the following: headache, unspecified pain, nausea and dizziness.
`Urinary incontinence is slightly less common, but apparently dose and drug—
`related as well. More serious, but much less common, adverse events 'seen
`at the same dose range have included vomiting, confusion, restlessness,
`agitation, somnolence and generalized weakness. No deaths that could be
`attributed to study drug have been reported at therapeutic doses of GHB
`One healthy 89 year old woman participating in pharmacokinetic trials
`developed dizziness, nausea, vomiting, respiratory depression and fecal
`incontinence, after a single (and initial) oral dose of 4.5 g of GHB.
`A single older narcoleptic patient who had been taking GHB for approximately
`1 1/2 years was hospitalized after an overdose of GHB 18 9. At the time of
`hospitalization he was comatose and unresponsive. He needed intubation
`
`

`

`Ranjit a. Mani, MD, HFD-120 Medical Review
`NDA 21196, Xyrem , Orphan Medical, Inc.
`
`Page 78 of 89
`,
`6/15/01
`
`and artificial ventilation, and awoke 6 hours later. This incident suggests that
`the safety margin between therapeutic and toxic doses may not be very wide
`. At therapeutic doses of GHB all adverse events appear to be reversible
`. While currently there is no strong evidence that GHB in therapeutic doses is
`epileptogenic or that episodes of urinary and fecal incontinence due to GHB
`are due to seizures, there is insufficient data at present to rule out either
`possibility.
`“Recreational” use of GHB, generally at doses, presumed or known to be
`higher than the therapeutic dose has been associated with adverse events
`that included fatalities attributable to the depressant effects of this drug on
`the nervous system. However concurrent use of alcohol and of other drugs
`with effects on the central nervous system has been reported in many of
`these instances
`
`.
`
`0 There is no evidence that GHB is toxic to any major organ other than the
`nervous system.
`
`14.2 Withdrawal Phenomena And Abuse Potential
`
`. There is no evidence from a small formal study with a randomized withdrawal
`paradigm (OMC-SXB-Zl) that the abrupt discontinuation of therapeutic doses
`of GHB used for 6 months to 3 V2 years leads to more than mild and _
`.
`infrequent withdrawal symptoms, except for a significantly increased '
`frequency of cataplexy.
`,
`. There are however a number of anecdotal reports of an actual withdrawal
`syndrome and, possibly, addiction in illicit “recreational” users of GHB, GBL
`or 1-4 80. In all these individuals, high doses of GHB or related drugs were
`believed to have been used at frequent intervals around-the-clock.
`
`‘
`
`14.3 Additional Comments Based On Review Of Major Amendment To NDA
`
`On March 23, 2001, the sponsor submitted a major amendment to this NDA
`
`'
`
`The purpose of the amendment was to address the following
`. Deficiencies in the open-label Scharf study outlined in the safety review
`. A number of questions pertaining to the safety data for clinical trials conducted by
`Orphan Medical
`Several related issues.
`
`.
`
`!
`
`In submitting the major amendment the sponsor requested a 90-day extension to the
`original Prescription Drug User Fee Act deadline of April 2, 2001.
`
`This major amendment is reviewed in a separate document. Please refer to that review
`for full details.
`
`‘
`
`A number of comments by me about the safety of Xyrem®, based on a review of the
`Amendment are below. In order to understand the context of the comments further, the
`reader will need to refer to the review of the Amendment itself which is in a separate
`document.
`V
`
`

`

`Ranjit B. Mani, MD, HFD-120 Medical Review
`NBA 21196, Xyrem , Orphan Medical, Inc.
`
`Page 79 of 89
`'~ 6/15/01
`
`The manner in which data for the Scharf study have been collected,
`recorded, and presented in this submission cannot be said to be ideal.
`Of the 80 patients who participated in the Scharf study and did not enter the
`currently ongoing Orphan Medical Treatment lND study OMC—SXB-7, 64
`patients might be stated to have be “accounted fof’ although the basis for
`doing so is less-than-optimal in a significant number. Further efforts need to
`be made by the sponsor to account fully for 11 of the remaining 16
`(unsuccessful recent efforts have been made to contact 5 patients out of
`those 16). The 11 patients are listed below. Adverse events that were
`ongoing at the time of discontinuation are reasons for obtaining further follow-
`up in at least some of these 11 patients
`
`01-004/r' “
`01-027/1/
`01 ~054/l __,
`01—065/:
`01-2w. ,
`01-24011
`“262/
`01-269/ J,
`01 283/ "
`01-268/ W,
`01 -256/
`
`None of the “adverse events” in the “unevaluable” category that occurred in
`the Scharf study appear to be attributable to GHB
`Urinary and fecal incontinence both appear to be unusually common a‘r'verse
`events in patients taking GHB and the key issues are whether such episodes
`are accompaniments of unrecognized convulsions, and whether GHB is
`capable of causing convulsions at therapeutic doses. Currently the evidence
`that the vast majority of episodes of incontinence in the entire NDA are
`related to unrecognized convulsions is weak. There does appear to be at
`least 1 patient in the Scharf study"in whom incontinence clearly accompanied
`a true convulsion.
`
`While there are clearly a few patients (n = 2) in the entire NDA safety
`database who experienced, or may have experienced, convulsions while
`taking GHB, the presence of confounding factors (e.g., possible
`benzodiazepine withdrawal) makes it difficult to link the convulsions causally
`to GHB. Whether GHB is capable of causing other types of seizures, e.g.,
`absence or partial complex, is even less clear
`In this NBA, and especially in the Scharf Study, the term “sleepwalking” has
`been used as a verbatim (investigator) term for a common adverse event.
`Detailed clinical descriptions of such episodes are not available for the
`majority of patients and their mechanism has not been delineated. A separate
`analysis of these episodes has not been performed by the sponsor and it is
`not clear how common they are in the Integrated Clinical Trials grouping, but
`such episodes have been associated with serious consequences (e. g.,
`overdose, pyrogenesis, consuming toxic chemicals)in patients enrolled in the
`Scharf study
`
`

`

`Ranjit e. Mani, MD, HFD-120 Medical Review
`NBA 21196, Xyrem . Orphan Medical, Inc.
`
`Page 80 of 89
`6/15/01
`
`The information available in this NDA does suggest that GHB is capable, at
`therapeutic doses, of causing a confusional state (which may be
`accompanied by psychotic symptoms). The incidence and seriousness of
`such adverse events may be slightly more pronounced at higher doses, and
`especially if higher doses are administered without titration. However a
`confusional state also appears to be capable of occurring at lower and even
`sub-therapeutic doses of GHB, and after maintenance treatment for several
`months. The presence of true confusion in patients taking GHB could lead to
`their taking GHB in a manner other than as prescribed. The symptoms that
`have been subsumed under the COSTART term “confusion” are not unusual
`
`for a sedative-hypnotic drug.
`in the majority of patients who developed “neuropsychiatric" adverse events
`(e.g., paranoia, hallucinations, anxiety, stupor, etc) while taking GHB in
`Integrated Clinical Trials it is not possible to attribute causality for the adverse
`event to GHB. Pre—existing psychiatric illness, and concomitant medications
`such as stimulants, as well as other factors, could be contributory. Even in
`patients in whom there was no recorded premorbid history of psychiatric
`illness the extent to which they were screened for such illness is not clear.
`However the occurrence of neuropsychiatric adverse events in patients taking
`GHB, even if not directly caused by the drug, could place them at risk of
`intentional or accidental overdose, as is suggested by the narratives in'this
`review
`,
`There is no firm evidence that any patients participating in the Integrated
`Clinical Trials had drug-induced lupus. However antinuclear antibody and
`antihistone antibody testing was not performed for patients participating in
`this study
`ThereIs no evidence suggesting a causal link to GHB for the small number of
`hypoglycemic an hyperglycemic blood test readings in the NDA; several of
`the apparently hypoglycemic readings could in fact have represented _,
`laboratory errors. Neither is there firm evidence in AERS or in the medical
`literature that GHB is capable of causing hypoglycemia.
`GHB is unlikely to have been the cause of transaminase elevations seen in a
`few patients in the integrated Clinical Trials.
`As noted above the manner in which the Scharf study was conducted was
`deficientIn many ways. Of particular concern was the lack of systematic,
`active surveillance for adverse events and missing drug accountability .
`records. As also noted earlier in this review (see Section15) the Centerfs
`Division of Scientific investigations is of the opinion that the Scharf Study.
`data are unacceptable and has recommended that this study not be used in
`support of the application. From this reviewer’s perspective the best that can
`be said about this study is that the vast majority of those enrolled have been
`“accounted” for in the sense that it is unlikely that they have suffered any
`catastrophic events that this Agency is unaware of. I would, therefore, ,
`,
`recommend that this study not be used in estimates of the adequacy of
`exposure to Xyrem® in the safety database (see next bullet)
`
`

`

`Ranjit B. Mani, MD, Him-120 Medical Review
`NBA 21196, Xyrem , Orphan Medical, Inc.
`
`Page 81 of 89
`6/15/01
`
`' The number of patients exposed to GHB in the NDA Safety Database minus
`the Scharf study appears sufficient to meet ICH guidelines at the 6-month
`and 1-year levels but not in regard to the total number of patients exposed;
`however allowance can be given for GHB being designated as an orphan
`drug and the total number exposed may therefore be acceptable.
`
`Note that the extent of exposure to GHB in patient-years is reduced by about
`79% once the Scharf study data are eliminated. Admittedly, the ICH
`guidelines do not specifically address the issue of desirable exposure in
`patient-years.
`'
`
`Further if one concludes (from the efficacy studies) that the 9 g/day dose is
`the only effective dose and is that to be recommended for general use the
`number of those exposed to Xyrem® at that dose for 2 6 months and 2 12
`months does not meet ICH guidelines. On the other hand if it is concluded
`from the efficacy studies that the effective dose of Xyrem® ranges from 6-9
`g/day, the number of those exposed at that dose range for 2 6 months falls
`somewhat short of ICH guidelines. Note that the sponsor has not supplied
`data for the total number of patients exposed for any duration (including or
`excluding the Scharf study) for the 6-9 g/day dose range or at the 9 g/day
`dose itself
`
`Note that ICH guideline E1A (July 1997) includes the following statements:
`
`“The number of patients treated for 6 months at dosage levels intended for clinical use should be adequate to
`characterize the pattern of adverse events over time. To achieve this objective the cohort of exposed SUDJeCiS
`should be large enough to observe whether more frequently occurring events increase or decrease over time as
`well as to observe delayed events of reasonable frequency (e.g., in the general range of 0.5% to 5%). Usually 300
`to 600 patients should be adequate...
`
`....... 100 patients exposed for a minimum of 1 year is considered to be acceptable to include as part of the safety
`database. The data should come from prospective studies appropriately designed to provide at least one year
`exposure at dosage levels intended for clinical use."
`
`14.4 Risk Management Program
`The small clinical trial safety database, the narrow margin of safety and the risks
`of abuse and misuse all call for approval to be conditional on a risk management
`system that is more stringent than that proposed by the sponsor. Key additional
`elements of such a system should include
`'
`Dispensing of the drug exclusively to patients with a diagnosis of cataplexy
`A,
`confirmed by their physicians
`Commitment by the sponsor to a detailed plan for active post-marketing surveillance
`for instances of diversion, abuse, misuse and adverse events of special concern
`Clear statements in the approved label, patient information sheet and patient and
`physician educational materials about the nature of the drug (i.e., that it contains the
`same active ingredient as illicitIy-used GHB), the limited experience with the drug
`during development, the potentially serious toxicity of both therapeutic doses ':.nd
`overdoses
`’
`
`

`

`Ranjit a. Mani, MD, HFo-izo Medical Review
`NDA 21196, Xyrem , Orphan Medical, inc.
`
`Page 82 of 89
`6/15/01
`
`0 Use of Subpart H of the Accelerated Approval regulations (21 CFR 314.500) so as to
`provide a means of restricting distribution of the drug and for enforcement of the risk
`management program. Justification for institution of these regulations is as follows
`0 Xyrem® is intended to treat a serious disease (cataplexy)
`'
`'
`o Xyrem® provides meaningful benefit to patients over existing treatment
`0 Xyrem® can be used safely only if its distribution or use is restricted
`
`15. Study Site Inspections
`
`15.1 Sites Inspected
`
`The following study sites pertinent to this application were inspected by the
`Division of Scientific lnvestiations DSI at the reuest of this Division
`
`Stud
`
`Orphan Medical, Inc
`Minnetonka, MN
`OMC—GHB-2
`
`
`(Sponsor)
`
` OMC-GHB-2
`
` OMC—GHB-2
`
`
`
`
`
`Oklahoma City, OK
`
`
`
`" Jonathan Schwartz, MD
`Investi-ator
`
`(So-onsor-Investiator
`
`S n-onsor-lnvestiator
`
`The results of these inspections are described in detail in a Clinical Inspection
`Summary written by Constance Lewin, MD, of the Division of Scientific
`,
`Investigations, dated June 11, 2001. Please refer to that document for full
`‘
`details.
`»
`
`These inspections uncovered a number of deficiencies, the most prominent of
`which pertained to the Scharf open-label study.
`
`15.2 Scharf Study Site Inspection
`
`This inspection is described in the NDA Safety Review
`
`15.3 Division Of Scientific Investigations Conclusions
`
`These conclusions are summarized in the followin table
`
`
`
`Orphan Medical, Inc
`(Snnnsnri
`
`
`
`
`
`Jonathan Schwartz, MD
`lnvesti-ator
`
`Lawrence Scrima, PhD
`(Sponsor—Investigator)
`
`
`
`
`Martin Scharf, PhD
`(Sponsor-Investigator)
`
`
`’Drug accountability records largely missing
`"Multiple deficiencies including missing drug accountability records
`
`Scharf Study
`
`
`
`
`Data acceptable
`
`
`
`
`
`Data unacceptable'.
`Recommendation: Stud not be used in su- oort of NDA
`
`Data unacceptable"
`.
`
`
`
`Recommendation: Stud not be used in sun . rt of NDA
`
`16. Financial Disclosure Certification
`
`Financial disclosure certification has been submitted with this application.
`
`

`

`Page 83 of 89
`Ranjit B. Mani, MD, HFD‘120 Medical Review
`
`NBA 21196. Xyrem , Orphan Medical, lnc. 6/15/01
`
`16.1 Components 0f Certification
`
`This certification has 2 components
`
`16.1.1 Certification Pertinent To Dr Lawrence Scrima
`
`The sponsor has supplied required financial disclosure information for Dr Scrima.
`
`Orphan Medical, Inc, entered into a financial contract with Dr Scrima on
`11/10/99. The contract allowed Orphan Medical to access documentation
`associated with the double-blind, placebo-controlled, cross-over trial in 20
`narcoleptic patients. The trial was conducted from April 5, 1986 to December 14,
`1987.
`
`The sponsor states that payments to Dr Scrima were made over 10 years after
`completion of the trial. While the payment was financially disclosable it did not
`have any impact on data collection, interpretation or analysis
`
`16.1.2 Certification Pertinent To Other Investigators
`
`The sponsor has supplied a list of 32 investigators who conducted clinical trials
`on behalf of Orphan Medical, Inc. In regard to this list the sponsor has
`. Certified that it has not entered into any financial agreement with the clinical
`investigators listed in the application whereby the compensation to the
`investigator could be affected by the outcome of the study in which the
`investigator was a participant, as defined by 21 CFR 54.2 (a)
`. Certified that each listed clinical inVestig