`
`
`
`
`
`
`
`CENTER FOR DRUG EVALUATION AND RESEARCH
`CENTER FOR DRUG EVALUATION AND RESEARCH
`
`
`
`
`
`APPLICATION NUMBER:
`22-192
`22-192
`
`
`APPLICA TION NUMBER:
`
`
`
`
`
`USUMMARY REVIEW
`
`SUMMARY REVIEW
`
`

`

`M E M O R A N D U M
`
`DEPARTNIENT OF HEALTH AND HUMAN SERVICES
`PUBLIC HEALTH SERVICE
`FOOD AND DRUG ADMINISTRATION
`CENTER FOR DRUG EVALUATION AND RESEARCH
`
`'DATE:
`
`March 27, 2009
`
`FROM:
`
`Thomas P. Laughren, MD.
`Director, Division of Psychiatry Products
`HFD-l30
`
`SUBJECT: Recommendation for approval action for iloperidone immediate release
`tablets for schizophrenia (for acute and maintenance treatment)
`
`TO:
`
`File NDA 22—192
`
`[Note: This overview should be filed with the 11—6-08 response to the
`agency’s 7-25-08 Not Approvable letter.]
`’
`
`1.0
`
`BACKGROUND
`
`It is an
`lloperidone is an atypical antipsychotic (5HT2 and D2 receptor antagonist).
`immediate release formulation for bid administration. This NDA seeks a claim for both
`
`the acute and maintenance treatment of schizophrenia, in a total dose range of 12 to 24
`mg/day. Iloperidone was developed under IND 36,827. This NDA was first submitted 9—
`27-07. We issued a Not Approvable letter on 7—25-08.
`There were two major
`deficiencies that were the basis for this action, i.e., (1) lack of sufficient effectiveness
`data, and (2) lack of sufficient safety data in a relevant dose range.
`In addition to these
`not approvable issues, there were four other issues noted in the letter: (1) data from Dr.
`Gilliam’s site; (2) need to repeat hepatic impairment study; (3) need for iloperidone and
`P-Gp interaction study; (4) need for safety update. We subsequently met with the
`sponsor on 9—10—08 (see meeting minutes) to discuss the Not Approvable action;
`
`2.0
`
`EFFICACY AND SAFETY DATA CONSIDERED IN ORIGINAL
`APPLICATION
`
`2.1
`
`Overview of Studies Pertinent to Efficacy
`
`The NDA contained 4 short-term (4 to 6-week), double-blind, randomized, parallel
`group, placebo-controlled trials in adult patients with acutely exacerbated schizophrenia
`or schizoaffective disorder (Studies 3101, 3005, 3004, and 3000). A114 studies involved
`fixed doses (or fixed dose ranges) for iloperidone, and all 4 had active controls. Three of
`the 4 studies included a mix of patients with schizophrenia and schizoaffective disorder.
`
`

`

`The sponsor also presented data from 3 longer-term trials (Studies 3001, 3002 and 3003)
`in support of a claim for maintenance efficacy in schizophrenia. The latter 3 studies were
`active controlled trials, comparing iloperidone with haloperidol, and found no differences
`between the 2 drugs. Since we have not accepted non-inferiority studies as a reliable
`source of evidence for efficacy claimsin schizophrenia, we did not comment further on
`these 3 studiesin the not approvable letter.
`
`2.2
`
`Basis for 7-25-05 Not Approvable Action (Lack of Sufficient Effectiveness
`Data)
`
`We accepted study 3101, a 4-week study comparing iloperidone 24 mg/day, ziprasidone
`160 mg/day, and placebo in acutely exacerbated schizophrenic patients, as a positive
`study. In our 7-25-08 not approvable letter, we expressed concerns about the remaining 3
`short-term studies, all in patients with a mix of schizophrenia and schizoaffective
`disorder. In our original review, we focused on the subsets of patients with schizophrenia
`in studies 3000, 3004, and 3005. Using this approach, we concluded that neither study
`3000 nor study 3004 provided evidence in support of a claim for efficacy in
`schizophrenia, while we considered study 3005 a possibly positive study in the
`schizophrenic subgroup in a dose range of 12-24 mg/day. We raised 2 additional
`concerns, however, that we considered sufficient at that time to not consider study 3005 a
`second source of evidence. The first concern was the relatively consistent finding that
`iloperidone appeared to be inferior to other treatments, across studies 3000, 3004, and
`3005. For study 3005, the iloperidone 12-16 mg/day vs risperidone 6-8 mg/day contrast
`favored risperidone (p=0.005), as did the iloperidone 20—24 mg/day vs risperidone 6-8
`mg/day contrast (p=0.093), albeit not at the usual p < 0.05 level of significance. A
`second concern was the observation in study 3005 that the positive effect for iloperidone
`over placebo was coming almost entirely from the non—US sites.
`
`2.3
`
`Basis for 7—25-05 Not Approvable Action (Lack of Sufficient Safety Data)
`
`We also noted in the not approvable letter our concerns about the prominent QT
`prolonging effect of iloperidone and the difficulty in titrating patients to an effective dose
`of iloperidone. We indicated that the QT signal would relegate iloperidone to essentially
`second line status. Based on the statistically significant superiority of risperidone 6-8
`mg/day to iloperidone 12-16 mg/day (p=0.005) in study 3005, we considered the
`iloperidone 20-24 mg/day dose range the only acceptable dose range for this drug in this
`study. Given that the only other source of positive evidence came from an iloperidone
`dose of 24 mg/day in study 3101, we raised a concern that the sponsor had safety data for
`Only 508 iloperidone patients in this dose range of 20-24 rug/day, including only 64
`patients treated for at least 6 months and only 22 for at least 1 year. Thus, we indicated
`that,» even if we were to accept the effectiveness data from studies 3101 and 3005 as
`sufficient, the sponsor would need at least 1000 additional patients exposed within the
`20-24 mg/day dose range, including 300 for 6 months and 100 for 1 year.
`
`

`

`2.4
`
`Summary of Efficacy Data for Studies 3000, 3004, and 3005
`
`The sponsor responded to our 7-25-08 not approvable letter with an initial 8-21-08
`response, and with several subsequent documents, and then requested a meeting with the
`division. We provided preliminary comments to the sponsor in which we expressed our
`continued concern that the application was deficient with regard to both efficacy and
`safety data. We did, however, acknowledge their complaint that they were not informed
`until the time of the action letter that we would focus on the subgroups of patients with
`schizophrenia in studies 3000, 3004, and 3005. We felt, however, that our advice to them
`to limit enrollment to patients with schizophrenia in study 3101 should have been a clear
`signal that this subgroup would be our focus in analyzing the other three studies as well.
`Nevertheless, we indicated that we would consider the data for both approaches, i.e., the
`schizophrenic subgroup and all patients randomized. What follows under this heading is
`the summary data for studies 3000, 3004 and 3005, using both approaches and the
`protocol specified analyses.
`[Notez these tables are taken from the final meeting minutes
`for our 9-10-08 meeting with the sponsor.]
`
`APPEARS THIS WAY
`0N ORIGINAL
`
`

`

`Study 3000
`
`FDA analysis: Table 1 summarizes the FDA’s analysis focusing on the schizophrenia sample. The primary
`contrast is between iloperidone 8mg and 12mg combined against placebo. The primary contrast did not
`separate from placebo (p=0.148), and therefore, no additional comparisons are permitted. Haloperidol is
`highly statistically significantly superior to placebo (p=0.005) and shows a numerical advantage over all
`three doses of iloperidone. Haloperidol is also numerically superior to iloperidone 8mg and 12mg
`combined, although this contrast just misses statistical significance (p=0.063).
`
`Table 1. Study ILP3000ST: FDA’s efficacy results: change from endpoint to baseline in PANSS
`total score OCF in the MI’I‘T sam 1 le excludin- schizoaffective - atients
`
`
`
`Sample size
`LS Means
`
`Difference from placebo
`Unadjusted p-values
`
`70
`
`78
`
`
`
`
`no 12 mg ' 110608+12mg Hal 15mg
`110 8 mg
`— 1130 4 mg
`
`
` Difference fi'om haloperidol
`
`Unad'usted' -values
`
`
`
`(Source: Vanda’s Meeting Package, Table 12, Page 27 and FDA’S results)
`
`Protocol-specified grimafl analysis: Table 2 summarizes the protocol-specified primary analysis that
`includes all randomized patients. The primary contrast is between iloperidone 8mg and 12mg combined
`against placebo. The primary contrast did not separate from placebo (p=0.065), and therefore, no
`additional comparisons are permitted. Haloperidol is highly statistically significantly superior to placebo
`(p<0.001) and shows a numerical advantage over all three doses of iloperidone. Haloperidol is also
`numerically superior to iloperidone 8mg and 12mg combined, and this contrast is now statistically
`significant (p=0.027).
`
`Table 2. Study ILP30005T: sponsor’5 primary efficacy results: change from endpoint to
`baselinein PANSS total score
`
`Sample size
`LS Means
`
`Difference from placebo
`Unadjusted p—values
`
`Difference from
`
`Haloperidol
`Unad‘usted . -va1ues
`
`(Source: Vanda’s Meeting Package, Table 14, Page 28 and FDA’s results)
`
`Comment: Thus, either approach to defining the sample for this study yields a negative result for
`iloperidone. With the sponsor’s preferred analysis including all randomized patients, the superiority of
`haloperidol over the primary iloperidone group (8 + 12 mg) is statistically significant. This study,
`therefore, provides no support for iloperidone but does suggest the statistically significant superiority of
`haloperidol over iloperidone
`
`

`

`Study 3004
`
`FDA analysis:
`
`Table 3 summarizes an analysis excluding schizoaffective patients. A sequential testing approach was
`employed. First, a comparison was carried out between the 10—16 mg/d group and the placebo group.
`Subsequently, iloperidone 4-8 mg/d was tested against placebo. The results suggest that both iloperidone
`groups did not separate from placebo. The results also suggest that risperidone was highly significant
`against placebo (p=0.001). A comparison between the two iloperidone dose groups against risperidone
`suggests that risperidone was superior to both iloperidone dose groups (p-value = 0.006 against iloperidone
`4-8 mg/d and p-value = 0.021 against iloperidone 10-16 mg/d).
`
`Table 3. Study ILP3004ST:CFDA’s efficacy results: change from endpoint to baseline in BPRS total
`score
`excludin_ schizoaffective atients; MITT samle
`
`
`
`1251
`110
`1196
`Sample size
`1185
`LS Means
`
`
`
`
`0.9
`Difference from placebo
`
`
`0.581
`
` Unadjusted p-values
`
`
`
`
`
` Difference from risperidone -4.5
`
`
`Unad'usted o-values
`0.006
`
`
`
`
`
`(Source: Vanda’s Meeting Package, Table 9, Page 23)
`
`Protocol-sgecified analysis: Table 4 summarizes the protocol-specified analysis that includes all patients
`(schizophrenia and sehizoaffective). Again, a sequential testing approach was employed. The comparison
`carried out between the 10-16 mg/d group and the placebo group was statistically significant (p—value =
`0.001) in favor of iloperidone 10-16 mg/d. Subsequently, iloperidone 4-8 mg/d was tested against placebo
`and was statistically significant (p-value = 0.012). A comparison between the two iloperidone dose groups
`against risperidone suggests that risperidone was superior to both iloperidone dose groups (p-value = 0.007
`against iloperidone 4-8 mg/d and p-value = 0.034 against iloperidone 10-16 mg/d).
`
`Table 4. Study ILP3004ST: sponsor’s primary efficacy results: change from endpoint to
`baseline'in BPRS total score
`0C in the MITT sam o le
`
`
`
`
`
`
`152
`149
`1423
`Sample size
`
`
`LS Means
`
`
`
` Difference from risperidone
`
`Unad'usted -values
`
`
`(Source: Reproduced from ILP30045t-1egacy Report; Table 9.1-2, page 543 and FDA’s results)
`
`Difference from placebo
`Unadjusted p-values
`
`Comment: Although the all-patients analysis yields a positive result vs placebo for both iloperidone dose
`groups, both analyses suggest clear inferiority of iloperidone at these doses to a standard dose range for
`risperidone. Thus, either approach to defining the sample for this study yields a result that favors a
`standard control agent over iloperidone.
`
`

`

`Study 3005
`
`FDA analysis: Table 5 summarizes an analysis excluding the schizoaffective patients. For this study, a sequential
`testing procedure was employed. Iloperidone 12-16 mg/d was tested first at a 0.05 level. If this test was significant
`then the iloperidone 20-24 mg/d would be tested. Both iloperidone dose groups were statistically significantly
`superior to placebo. The results also suggest that risperidone was numerically, if not statistically, superior to
`iloperidone at the 20-24 mg/day group (p=0.093), and both numerically and statistically significantly superior to
`iloperidone at the 12—16 mg/day dose (p=0 .005).
`
`,
`
`Table 5. Study ILP30058T: FDA’s efficacy results: change from endpoint to baseline in BPRS total
`score OCF excludin schizoaffective a atients,- MITT sam 1e
`
`
`
`
`
`
`1181
`178
`Sample size
`119
`113
`LS Means *
`11.4
`
`
`71
`Difference fi‘om placebo
`
`
`
`<0.001
`Unadjusted p-values
`
` Difference from risperidone
`
`Unad'usted . values
`
`
`
`(Source: Vanda’s Meeting Package, Table 5_, page 18 and FDA’s results)
`
`Protocol-specified analysis: Table 6 summarizes the protocol-specified primary analysis including both
`schizophrenia and schizoaffective patients. Iloperidone 12—16 mg/day did not separate from placebo (p-value =
`0.09). Consequently, iloperidone 20—24 mg/d cannot be considered. We concluded this was a negative study based
`on the primary analysis. Risperidone appears to be superior to both iloperidone 12-16 mg/d and 20-24 mg/d (p-
`values < 0.001 and 0.034, respectively).
`
`Table 6. Study ILP3005ST: sponsor’s primary efficacy results: change from endpoint to baseline in
`BPRS total score LOC in the MITT sam 1e
`
`
`
`
`Sample size
`LS Means*
`
`2310
`
`
`
`Difi‘erence from placebo
`
`
`Unadjusted p-values
`
`
`
`
`
`
`1111461
`
`152
`
`148
`11.5
`6.5
`<0.001
`
`Difference from risperidone
`Unad'usted -values
`
`
`(Source: Reproduced fiom ILP3005st—legacy Report; Table 9.1-2, page 586 and FDA’s results)
`
`Comment: For this study, the all-patients analysis yields a negative result for the 12-16 mg/day group, and,
`therefore, the 20-24 mg/day group cannot be considered. Furthermore, this is yet another demonstration of the
`apparent inferiority of iloperidone at these doses to a standard dose range for risperidone. Thus, once again, either
`approach to defining the sample for this study yields a result that appears to favor a standard control agent over
`iloperidone.
`
`In summary:
`0 For study 3000, whether the analysis focuses on the schizophrenic subgroup or all patients, it
`does not provide evidence of efficacy for iloperidone. Furthermore, in the sponsor’s
`preferred analysis including all patients, haloperidol is clearly superior to placebo and
`appears to be statistically significantly superior to iloperidone.
`
`

`

`o For study 3004, the analysis including all patients does show superiority of iloperidone over
`placebo, a finding that is not seen for the analysis including only schizophrenic patients. In
`both instances, however, risperidone appears to be statistically significantly superior to
`iloperidone.
`- For study 3005, only the analysis focused on the schizophrenic subgroup shows superiority
`of iloperidone over placebo. In the sponsor’s preferred analysis, iloperidone fails to show
`superiority to placebo and, at the same time, risperidone appears to be statistically
`significantly superior to iloperidone.
`
`2.5
`
`Comment on 9-10-08 Meeting with Sponsor
`
`I refer to the final minutes for this meeting for complete details on the proceedings and outcome.
`
`The sponsor made a number of arguments in favor of iloperidone, several of which we
`acknowledged had some plausibility and encouraged them to elaborate on in their complete
`response: (1) their contention that the apparent inferiority of iloperidone to the active
`comparators risperidone and haloperidol is only temporary due to differences in the time it takes
`to get patients to effective exposures for iloperidone, (2) their contention that this early
`difference does not represent a significant risk to patients, and (3) their contention that there are
`certain safety advantages that iloperidone has over other antipsychotic drugs in the class that
`tend to mitigate this early disadvantage in efficacy. We did not accept their argument about the
`need for multiplicity adjustments for the comparisons of active control drugs to iloperidone.
`Using a sequential approach, there still remain 2 illustrations of an apparent disadvantage for
`iloperidone in efficacy.
`
`There remained a concern, however, about the primary source of evidence for the efficacy of
`iloperidone.
`-We still considered study 3101 the only unambiguously positive study.
`—We remained concerned about the sponsor’s focus on study 3004 as a primary source of
`evidence, since this study was based on the all-randomized patients analysis. We found this
`problematic because of the striking differences in outcomes for the schizophrenic and
`schizoaffective subgroups, and the analysis focusing only on the schizophrenic subgroup is not
`positive.
`,
`-We continued to consider study 3000 negative overall.
`-We continued to consider study 3005 as a possible primary source of support. We favored the
`analysis focusing on the schizophrenic patients, which we found to be positive. We remained
`concerned, however, about the geographic disparity in results. The positive findings came
`almost entirely from non-US sites. Thus, we indicated that they would need to make a '
`convincing argument that this disparity should not be a concern.
`
`3.0
`
`SPONSOR’S 11-6—08 RESPONSE TO 7-25-08 NOT APPROVABLE LETTER
`
`In its 11-6-08 response, the sponsor argues that they have provided positive results for the
`effectiveness of iloperidone in the acute treatment of schizophrenia in 2 adequate and well-
`controlled trials, i.e., studies 3101 and 3004. They further argue that studies 3000 and 3005
`
`

`

`provide supportive evidence for the acute efficacy of iloperidone and that- studies 3001, 3002,
`and 3003 provide evidence for the maintenance efficacy of iloperidone in schizophrenia. They
`acknowledge our arguments that they have not provided sufficient evidence for the acute and
`maintenance efficacy of iloperidone in schizophrenia, but note that they disagree. They indicate
`that they can show that iloperidone is effective for this indication in the US population, has
`comparable efficacy to other available antipsychotic agents, and has certain safety advantages
`over other available antipsychotic agents.
`
`3.1
`
`Evidence of Efficacy for Iloperidone
`
`As noted, the sponsor considers studies 3101 and 3004 as their primary sources of support for
`the efficacy of iloperidone in the acute treatment of schizophrenia, and studies 3000 and 3005 as
`supportive sources of evidence. Since we agree that study 3101 is a positive study, I will not
`comment further on that study (already fully described in previous reviews).
`I will also not
`comment further on studies 3001, 3002, and 3003. These were active-controlled maintenance
`studies that relied on findings of no difference between iloperidone and haloperidol. We have
`not yet accepted non-inferiority designs as a primary source of support for efficacy claims.
`
`'
`
`3.1.1 Study 3004
`
`The results for study 3004 are provided in section 2.4, but as noted in section 2.5, we still do not
`find this study an acceptable source of evidence. Although the results for both dose groups in
`the all-randomized patients analysis are positive, there is a striking difference in outcomes for
`the schizophrenic and schizoaffective subgroups.
`The analysis
`focusing only on the
`schizophrenic subgroup is not even close to positive for either dose group (p=0.306 for 10—16
`mg/day and p=0.581 for 4-8 mg/day).
`'
`
`In its response, the sponsor focuses heavily on an argument that we should not distinguish
`between schizophrenia and schizoaffective disorder. They argue that this is a difficult distinction
`in the acute setting and that the treatment of psychotic symptoms in the acute setting is the same,
`regardless of the diagnosis.
`'
`Comment: While these may be true statements, they are irrelevant to this discussion. The
`investigators had some basis for making the diagnoses they made, and it is highly
`inappropriate for the sponsor or its consultants to discard these judgments that the
`protocol required the investigators to make. The more important facts are that these are
`considered distinct diagnoses by DSM—IV, however difficult it is to distinguish them
`acutely, and the outcome is strikingly different for the different subgroups.
`
`The sponsor also argues that the finding for the schizoaffective subgroup, making up a mere 22%
`of the sample, is an “anomaly.” They note that the schizoaffective patients on placebo had
`scores that actually worsened by an average of 5.9 points compared to an improvement of 4.9
`points in the schizophrenic placebo sample.
`Comment: This may be an anomaly, however, the problem is not fixed by simply
`combining all the data to reach a conclusion. The anomaly is what is driving the overall
`positive result. Given this finding, the entire result should be discarded. This is clearly
`not just a power problem. We would not have objected if the effects were similar in the
`
`

`

`schizophrenic and schizoaffective patients, and the p-value for the schizophrenic
`subgroup would become significant
`if the sample size had been increased slightly.
`Clearly that would not have helped in this situation. Consequently, I am not persuaded
`by Vanda’s arguments for relying on the all patients analysis for study 3004.
`I do not
`consider study 3004 a valid source of evidence to support a claim for the acute treatment
`of schizophrenia.
`'
`
`3.1.2 Study 3000
`
`The sponsor continues to argue that the finding for the 12 mg/day dose arm for the all patients
`analysis (p=0.047) should be considered supportive evidence.
`Comment: As noted under section 2.4, this is a negative study whether considered for the
`schizophrenic subgroup or the all patients analysis. One is not entitled to look at the 12 '
`mg/day group for the all patients analysis because the combined 8+12 mg/day group is
`not significant. Thus, this study cannot be considered a primary source of support for the
`efficacy of iloperidone. Nevertheless, I do agree that the data for the 12 mg dose are
`consistent with the positive finding in study 3005 for efficacy in a range of 12 to 16
`mg/day.
`
`3.1.3 Study 3005
`
`The sponsor concedes that study 3005 is a negative study based on their preferred all patients
`analysis. Based on several other findings, they feel it can still 'be considered a source of
`supportive evidence.
`In our not approvable letter, we focused on analyses looking at the
`schizophrenic subgroups in studies 3000, 3004, and 3005. In these subgroup analyses, we found
`study 3005 to be a positive study (see section 2.4). We did, however, raise 2 concerns about this
`study. One concern had to do with the apparent inferiority of iloperidone to risperidone in this
`study. The other concern was the fact that the positive findings appeared to be coming almost
`entirely from the non-US sites. The issue of comparative efficacy will be addressed in section
`3.2.
`In responding to our concern about the geographic disparity, the sponsor used several
`arguments:
`(1) First, they argued that, in studies 3101 and 3004, subgroup analyses in the US population
`demonstrated a consistent superiority for iloperidone over placebo.
`(2) They also argue that, for the US subgroup in study 3005, neither iloperidone nor risperidone
`was superior to placebo. Thus, they consider the US subgroup in this study a failed experiment
`that cannot be interpreted. The following table supporting this assertion was taken from the
`statistics
`’
`review:
`
`

`

`Table 36. Study 1LP3OOSST: ret-‘iewer’s primary efficacy results by region: change from
`endpoint to baseline in BPRS total score (LOCI?) (excluding schizoaffecfive patients); MITT
`_
`sample
`'
`
`110 20-24 mg Risp 6-8 mg
`110 10-16 mg
`
`
`LISA.
`
`
`50
`48
`Sample size
`75
`
`
`
`6.53
`6.66
`8.99
`L8 Means 8
`
`
`
`
`1.08
`1.21
`Diflerence. from placebo
`3.54
`
`
`
`
`(-3 .06, 5.22)
`(3.32. 5.74)
`{-0.98, 8.07)
`(95% confidence intewaI)
`
`
`
`
`0.60?
`0 .599
`0.124
`Unadjusted p—values
`
`Nan-USA.
`
`Sample size
`103
`61
`71
`
`
`
`
`L8 Means 3
`8.33
`10.56
`13.28
`
`
`
`
`Difference from placebo
`4.89
`7.1 1
`9.84
`
`
`
`
`(5.65, 14.02)
`(95% confidence interval)
`(1.02, 8-25)
`(2.80, 11.42)
`
`
`
`
`
`
`0.001 {0.001
`
`0014
`Unad'usted -values
`(S ource: reviewer’s results)
`* Reviewer’s note: positive changes indicate improvements
`
`
`
`
`
`
`(3) Finally, they argue that the non-US sites for study 3005 (i.e., Canada, Germany, Hungary,
`Poland, Croatia, and Israel) all represent ethnic groups that are also well-represented in the US
`and, thus, can be considered reasonable surrogates for the US population.
`Comment:
`The geographic disparity was never a major concern for me, and I still
`consider it reasonable to look primarily at the schizophrenic subgroup as we did in our
`original analysis of this study.
`I accept their argument that the data from the US sites are
`not interpretable, as ‘no effect is shown for either active drug. Study 3101 does provide
`evidence of efficacy in a US population. Thus, I consider study 3005 a second primary
`source of evidence for the claim of the acute treatment of schizophrenia.
`I will address
`the issue of comparative efficacy in section 3.2.
`
`Efficacy Summary
`
`I' am in agreement with the sponsor that study 3101 is one primary source of positive evidence
`for iloperidone.
`I disagree with them, however, on study 3004' as a second primary source of
`evidence.
`I think the data for that trial are fatally pathological, and one cannot reasonably pool
`data from the schizophrenic and schizoaffective subgroups. It is not thatI fundamentally object
`to pooling data from schizophrenic and schizoaffective patients (we have accepted this approach
`many times in the past), but for this study, where the positive finding in the schizoaffective
`patients is by the sponsor’s own admission an “anomaly,” there is no justification for such a
`pooling. Thus, I find this study uninterpretable. Study 3000 also cannot be a primary source of
`support for iloperidone (but the data for 12 mg are consistent with findings in the 12-16 mg
`range for study 3005). That leaves only study 3005. As I have indicated, I am now willing to
`consider study 3005 as a second primary source of evidence for iloperidone. As noted, the 2
`objections to the data for this trial were the comparative efficacy findings, which I will address
`in section 3.2, and the geographic disparity.
`I am persuaded by the sponsor’s argument that the
`lack of efficacy in the US sites for study 3005 should not rule out this study as a source of '
`evidence. As they point out, risperidone also failed in the US sites, and thus, data from these
`sites is simply uninterpretable. They have other positive data for US sites, in particular, from
`
`°
`
`10
`
`

`

`study 3101. Although I do not believe we should rely on the noninferiority findings from the 3
`maintenance studies (studies 3001, 3002, and 3003) as support for a maintenance claim for
`iloperidone,
`I do believe this additional suggestive evidence of iloperidone’s efficacy in
`schizophrenia can be considered in the overall decision for this drug. Thus, I believe the sponsor
`has provided data from 2 adequate and well—controlled trials in support of a claim for the acute
`treatment of schizophrenia.
`'
`
`3.2
`
`Relative Efficacy of Iloperidone vs Other Antipsychotic Drugs
`
`As noted, one of our concerns expressed in the not approvable letter was the rather consistent
`finding across studies 3000, 3004, and 3005 that iloperidone appeared to be inferior on the
`primary efficacy assessment to an active control antipsychotic agent. This was true whether or
`not the study could be considered positive for iloperidone (see section 2.4). The reason we
`considered this problematic was that such inferiority might represent a risk to schizophrenic
`patients. In this section, I will consider the sponsor’s arguments about comparative efficacy, and
`in section 3.3 I will consider the sponsor’s arguments that, even if true, such an inferiority could
`not reasonably be considered a risk to patients.
`
`The sponsor proposes 4 arguments for why we should not be concerned about the finding of
`apparent inferiority of iloperidone to an active control agent in studies 3000, 3004, and 3005:
`
`-Study 3101 (iloperidone vs ziprasidonez: The sponsor points out that iloperidone was shown to
`be equivalent to ziprasidonein study 3101, a placebo-controlled short-term study.
`Comment: I agree that this was the case. However, this was the only one of the 4 short—
`term placebo controlled trials in which this was the case.
`'
`'
`
`- 2 ngday iloperidone vs haloperidol in studies 3000, 3001, 3002, and 3003: The sponsor
`points out that the 12 mg/d‘ay dose group in study 3000 was not'statistically significantly
`different than the haloperido] arm in that study (p=0.4). They also note that in the acute phase of
`the 3 haloperidol controlled trials (i.e., no placebo group; studies 3001, 3002, and 3003),
`iloperidone 12 mg/day was equivalent to haloperidol.
`‘
`Comment: Again, I don’t disagree with this observation. However, the active control
`studies (3001, 3002, and 3003) are difficult to interpret. And the 12 mg/day iloperidone
`arm is only 1 of several iloperidone dose groups in study 3000. The other iloperidone
`dose groups were inferior to haloperidol. On the other hand, it is true that 12 mg/day
`appears to/be the lower end of the effective dose range for iloperidone.
`
`-Apparent Inferiorifl as a Result of Titration Differences: The sponsor argues that the apparent
`inferiority of iloperidone to active control agents in several of their studies can be largely
`explained by slower titration rates for iloperidone compared to the active control agents in those
`trials. The result of this slower titration was a higher dropout rate for lack of efficacy in the
`iloperidone arms.
`[For example, in study 3005, dropouts for lack of efficacy were 23% in both
`the 12-16 and the 20-24 mg/day iloperidone arms, compared to 8% for the. risperidone arm.]
`In
`the LOCF analyses, these higher dropouts resulted in an apparent inferiority. The sponsor argues
`that, when titration for the control agent was done at a slower, comparable rate to that for
`iloperidone, efficacy was comparable for iloperidone and the active control agent. The sponsor ,
`
`11
`
`

`

`cites studies 3101 (a comparison with ziprasidone) and the 12 mg iloperidone vs 15 mg
`haloperidol comparison in study 3000 as examples of comparable titration Where efficacy was
`also comparable.
`In both studies, the titration period to achieve the target dose was 7 days for
`both active agents.
`
`The 2 more problematic studies for iloperidone, in terms of comparative efficacy, were studies
`3004 and 3005. In both studies, iloperidone appeared to be inferior to risperidone with regard to
`the efficacy outcomes. Given my current view that study 3004 has a pathological outcome that
`renders it completely uninterpretable,
`I will not consider it further in this discussion of
`comparative efficacy. That leaves only study 3005.
`It is true that the time periods to reach an
`effective target dose of iloperidone and risperidone were strikingly different in study 3005, i.e., 7
`days for iloperidone and 2 days for risperidone. Because it takes roughly 5 days to reach steady
`state concentrations with iloperidone. once the target'dose is achieved, it would be roughly 2 ,
`weeks before patients could achieve steady state concentrations. The sponsor conducted
`exploratory analyses looking at mean change from baseline to endpoint for the subgroup of
`patients who remained on assigned treatment for at least 2 weeks, in order to assess how the
`different active treatments might compare for patients who could achieve an effective steady
`state concentrations for both active drugs. It is of some interest that this analysis, although post-
`hoc and clearly exploratory, does show comparable efficacy for both iloperidone dose groups
`(10—16 and 20-24 mg/day) and risperidone 6—8 mg/day. The following figures (taken from
`sponsor’s 11-6-08 response to the NA letter) illustrate this finding of comparable efficacy for
`iloperidone and risperidone in study 3005 for patients remaining in treatment for at least 2
`weeks, and also in the observed cases analysis for this study:
`
`Figure 11:
`
`Study 3005: Mean Change from Baseline in 18—Item BPRS Score - '92 ‘Week
`LOCF and 0C Analysis
`
`Patienb Treated >2 Weeks
`
`Observed Cases
`
`
`5
`
`5
`
` 4
`
`+m""““"'”'“"‘"‘"l
`+lhpoMlm2l-I6mquw
`“mm-u"... I-..”
`mega
`‘=p-nlu$mumn
`
`L M
`
`'1‘
`
`I: D Ghana:
`
`
`184nmMR8scone(LOCFI b
`
`
`
`fromalaninein18-h"-EPRSSaute{DC}
`Change
`(mmBawtlnetn
`
`0'1, '
`
`.14
`
`.18
`
`-c-lwn:nnnl 1248me
`+Iwrmm-maamwu
`+nnumon-osno'on
`,m
`i=p—nmasmulmo
`
`.__~'
`
`“'3 __ __ n
`'
`
`Source: Module 2.7.3 Figure 11 and primary data submitted in NDA 22—192
`
`It is noteworthy that our MMRM analysis (an analysis that is p