CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`
`APPLICATION NUMBER:
`
`022567Orig1s000
`
`SUMMARY REVIEW
`
`
`
`
`

`

`DEPARTMENT OF HEALTH AND HUMAN SERVICES
`
`
` PUBLIC HEALTH SERVICE
`
`
`FOOD AND DRUG ADMINISTRATION
`CENTER FOR DRUG EVALUATION AND RESEARCH
`
`January 18, 2011
`
`Thomas P. Laughren, M.D.
`Director, Division of Psychiatry Products
`HFD-130
`
`M E M O R A N D U M
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`DATE:
`
`FROM:
`
`
`
`
`
`SUBJECT: Recommendation for approval action for vilazodone tablets as a treatment for
`major depressive disorder (MDD).
`
`File NDA 22-567
`[Note: This overview should be filed with the 3-22-10 original submission of this
`NDA.]
`
`
`TO:
`
`
`
`BACKGROUND
`
`
`
`1.0
`
`Vilazodone is a new antidepressant that has been developed for the treatment of MDD.
`Vilazodone’s antidepressant effect is thought to be mediated through its activity as an SSRI and
`as a partial agonist at the 5HT1A receptor. There are multiple other drugs in the antidepressant
`class already approved for the treatment of MDD, however, this would be the first with this
`particular combined activity. This application is based on data from 2 short-term trials. The
`proposed dose is 40 mg/day, to be given on a qd basis, with food. Vilazodone is available as 10,
`20, and 40 mg immediate release tablets.
`
`The studies in support of this application were conducted under IND 54613. Several meetings
`were held with the various sponsors over the course of its development. The IND has been held
`by several different sponsors over its long course. The IND was initially submitted 11-21-97.
`An early EOP2 meeting was held with the sponsor on 12-20-05, however, this was clearly
`premature. This meeting focused on a proposed design for the first phase 3 study. Subsequent
`meetings were held on 8-7-06, 8-20-07, and 5-11-09. The 8-7-06 meeting was another early
`meeting to discuss CMP, OCP, and pharm/tox issues. The 8-20-07 meeting was focused on their
`genetic analysis plan, since at that time the sponsor hoped to utilize biomarkers in their
`development program. The 5-11-09 meeting was to discuss their second phase 3 study and again
`their plans for genetic analyses. A preNDA meeting was planned for June, 2009, however, this
`was cancelled since the sponsor found our preliminary comments sufficient to answer all of their
`questions. Ultimately, they dropped their plans to include analyses of genetic data as part of this
`NDA.
`
`
`
`Reference ID: 2892948
`
`1
`
`

`

`CHEMISTRY
`
`PHARMACOLOGY
`
`The primary clinical reviewer for this application was Dr. Cheri Lindberg and the primary
`statistical reviewer was Dr. Philip Dinh. A secondary review of this application was conducted
`by Dr. Robert Levin.
`
`
`2.0
`
`The CMC review was conducted by Drs. Pei-I Chu, Ph.D. and Tien-Mien Chen, Ph.D., and they
`have recommended approval. Rik Lostritto, Ph.D., has written a Division Director memo
`confirming that all CMC issues have been resolved, and has also recommended an approval
`action. The preapproval inspections have been satisfactorily completed. The proposed name,
`Viibryd, has been accepted by DMEPA.
`
`
`3.0
`
`The pharm/tox review was conducted by Violetta Klimek, Ph.D. and supervisory overviews were
`provided by Linda Fossom, Ph.D., Barry Rosloff, Ph.D, and Paul Brown, Ph.D.. All pharm/tox
`questions and issues have been resolved, including agreement on the pharm/tox sections of final
`labeling.
`
`
`
`
`
`
`
`
`Nevertheless, they have now finally accepted our proposed language for mechanism of action in
`section 12.1.
`
` The pharm/tox group does
`not have any other concerns that would preclude a final approval action for this application.
`
`The 2 major human metabolites of vilazodone (M10 and M17) do not appear to have important
`serotonergic activity. Both have been adequately assessed for toxicity, however, it is unclear if
`one (M17) has been assessed for embryofetal toxicity, since its presence was not confirmed in
`the rat or rabbit studies. The sponsor has agreed to explore this issue post-approval.
`
`Specifications for several genotoxic or potentially genotoxic impurities have been limited so that
`human exposure will be no more than
` µg of each per day at the MRHD of 40 mg/day.
`
`The sponsor has agreed to conduct a juvenile animal study in rats prior to conducting pediatric
`studies in children less than 13 years of age.
`
`Teratogenicity, carcinogenicity, mutagenicity, and fertility findings with vilazodone can be
`summarized as follows:
`
`
`
`
`Reference ID: 2892948
`
`2
`
`(b) (4)
`
`(b) (4)
`
`(b) (4)
`
`

`

`BIOPHARMACEUTICS
`
`-Vilazodone caused some developmental toxicity in rats and rabbits at doses that were several
`multiples of the maximum recommended human dose (MHRD), but not at lower multiples, i.e.,
`10 times the MRHD for rats and 4 times the MRHD for rabbits. Vilazodone was not teratogenic
`in either species.
`-Two-year carcinogenicity studies were conducted in B6C3F1 mice and Wistar rats. There were
`no findings in the rat study. In mice, hepatocellular carcinomas and mammary gland
`adenocarcinomas were observed at doses that were several multiples of the maximum
`recommended human dose (MHRD), but not at lower multiples, i.e., 5.5 times the MRHD for the
`hepatocellular carcinomas and 1.8 times the MRHD for the mammary gland adenocarcinomas.
`The clinical significance of these findings for humans is unknown.
`-Mutagenicity assays were mixed, with negative findings in two in vitro assays, but positive
`findings in two in vitro clastogenicity assays. Findings were negative, however, in two in vivo
`clastogenicity assays, and in an in vivo/in vitro unscheduled DNA synthesis assay.
`-A fertility assay in rats revealed positive findings in males at 40 times the MRHD, but not at 6
`times the MHRD.
`
`
`4.0
`
`The OCP review was conducted by Drs. Bei Yu, Huixia Zhang, Jee Eun Lee, Atul Bhattaram,
`Yaning Wang, Issam Zineh, and Li Zhang.
`
`There were 24 phase 1 studies in the vilazodone program, including 9
`bioavailability/bioequivalence studies, a mass balance study, 2 food effect studies, renal and
`hepatic impairment studies, an elderly study, 2 drug-drug interaction studies, and 7 special
`studies. The 7 special studies included: PET study; REM suppression study; thorough QT study;
`ethanol interaction study; gastric pH study; a sleep EEG study; and a study on ejaculatory
`effects. There were also 9 in vitro studies.
`
`PET Study
`
`Study 255 was conducted to evaluate occupancy for the 5HT1A receptor. Single vilazodone
`doses of 20 and 40 mg were assessed. There was little evidence for occupancy at the 20 mg
`dose, however, 5HT1A occupancy at the 40 mg dose was found to be in the range of 15-35%.
`This study was considered part of the basis for the 40 mg dose selection for the definitive phase
`III trials. It might be argued, however, that multiple dose studies would have been preferable,
`given the relatively long elimination half-life of vilazodone (around 26 hours) with an
`accumulation ratio of approximately 1.8.
`
`Pharmacokinetic Profile for Vilazodone
`
`Vilazodone’s clinical effects are thought to be due primarily to the parent drug. Its
`pharmacokinetic properties are summarized in the following table, from the OCP review:
`
`
`Reference ID: 2892948
`
`3
`
`

`

`Absorption
`
`Distribution
`
`Metabolism
`
`Excretion
`
`Food Effect
`
`Protein Binding, %
`
`Pathways
`
`High fat/light meal
`increased Cmax and AUC by
`~2-fold.
`96-99
`CYP (3A4 is the primary
`isoenzyme, with the minor
`contributions from
`CYP2C19 and 2D6) and
`non-CYP (possibly by
`carboxylesterase). No
`active metabolites.
`A mass balance study for vilazodone showed ~85% of the
`administered radioactivity was recovered in the urine
`(~20%) and feces (~65%) combined, while ~ 3% of the
`administered dose of vilazodone was recovered as
`unchanged drug (~1% in urine, and ~2% in feces).
`
`Table 1: Important PK properties of vilazodone
`
`
`PK Property
`Dose-proportionality
`
`PK Parameter
`PK dose-proportional for doses 5 – 80 mg
`Tmax (median), hour
`4-5
`T1/2, hour
`~25
`Absolute Bioavailability
`72
`(with food), %
`
`
`Accumulation of vilazodone is predictable from single dose data (accumulation factor of about
`1.8), does not vary with dose, and steady state is achieved in about 3 days. The steady state
`mean Cmax value after daily dosing with vilazodone 40 mg under fed conditions is 156 ng/mL.
`
`The following figure from the OCP review illustrates the time-concentration profile for
`vilazodone following a single 40 mg dose:
`
`
`Reference ID: 2892948
`
`4
`
`

`

`Figure 1. Mean (± SD) vilazodone plasma concentration versus time profile after oral single
`dose of 40 mg vilazodone tablet under fed conditions.
`
`20
`
`40
`
`60
`
`80
`
`100
`
`120
`
`140
`
`160
`
`140
`
`120
`
`100
`
`80
`
`60
`
`40
`
`20
`
`0
`
`Vilazodone Plasma Concentration
`
`(ng/mL)
`
`0
`
`-20
`
`Time (h)
`
`
`The Effect of Intrinsic and Extrinsic Factors on Vilazodone’s Pharmacokinetics
`
`Effect of Intrinsic Factors: The study of various intrinsic factors on vilazodone’s
`pharmacokinetics suggest that no dosage adjustment is needed based on age, gender, renal, or
`hepatic impairment (See Figure 2, from the OCP review). Patients with severe renal or hepatic
`impairment have not been studied. Given the mechanisms for clearance of vilazodone, a study in
`patients with severe renal impairment would not be needed, however, it would be informative to
`have a study in patients with severe hepatic impairment.
`
`
`
`
`Reference ID: 2892948
`
`5
`
`

`

`Figure 2. Impact of intrinsic factors on vilazodone PK
`
`
`Population Description
`
`PK
`
`Fold Change and 90%CI
`
`Recommendation
`
`Age:
`>65 years
`
`Gender:
`Females
`
`Renal Impairment:
`Mild
`
`Moderate
`
`Severe
`
`Hepatic Impairment:
`Mild
`
`Moderate
`
`Severe
`
`Cmax
`AUC
`
`Cmax
`AUC
`
`Cmax
`AUC
`Cmax
`AUC
`Cmax
`AUC
`
`Cmax
`AUC
`Cmax
`AUC
`Cmax
`AUC
`
`No dose adjustment
`
`No dose adjustment
`
`No dose adjustment
`
`No dose adjustment
`
`No dose adjustment
`
`No dose adjustment
`
`No dose adjustment
`
`Not studied
`
`1.8
`1.4
`1.0
`0.6
`Change relative to reference
`
`The data shown for elderly subjects (>65 years) are relative to subjects (24-55 years).
`The data shown for female subjects are relative to male subjects.
`The data shown for renal and hepatic impairment are relative to subjects with normal
`renal and hepatic function, respectively.
`
`
`
`Effect of Other Drugs: The study of several drugs in combination with vilazodone revealed a
`concern for pharmacokinetic interactions only with strong 3A4 inhibitors (see Figure 3 to follow;
`from OCP review). Based on the finding with ketoconazole, the vilazodone dose should be
`reduced to 20 mg when used in combination with CYP3A4 strong inhibitors. Although the
`interaction of vilazodone with CYP3A4 inducers has not been evaluated, it can be expected that
`such inducers could result in inadequate vilazodone concentrations and may diminish
`effectiveness. Concomitant administration of vilazodone with inhibitors of CYP2C19 and
`CYP2D6 is not expected to alter plasma concentrations of vilazodone. These isoforms are minor
`elimination pathways in the metabolism of vilazodone. In vitro studies have shown that
`CYP1A2, CYP2A6, CYP2C9 and CYP2E1 have minimal contribution to the metabolism of
`vilazodone.
`
`
`Reference ID: 2892948
`
`6
`
`

`

`Figure 3. Impact of other drugs on vilazodone PK
`
`
`Change due to
`
`PK
`
`Fold Change and 90% CI
`
`Recommendation
`
`Ethanol:
`
`CYP3A4 Inh bitors:
`Ketoconazole
`
`Proton Pump Inhibitors:
`Pantoprazole
`
`Cmax
`AUC
`
`Cmax
`AUC
`
`Cmax
`AUC
`
`No dose adjustment
`
`Maximum dose: 20 mg
`
`No dose adjustment
`
`2.00
`1.50
`1.00
`0.50
`Change relative to reference
`
`
`
`
`Effect of Vilazodone on Other Drugs: Coadministration of vilazodone with substrates for
`CYP1A2, CYP2C9, CYP3A4, or CYP2D6 is unlikely to result in clinically significant changes
`in the concentrations of the CYP substrates. A study in healthy subjects found that vilazodone
`(20 mg/day for 8-10 days) had no effect on the pharmacokinetics of caffeine, flurbiprofen,
`nifedipine or debrisoquine, probes for CYP1A2, CYP2C9, CYP3A4, and CYP2D6, respectively.
`Vilazodone coadministration with mephenytoin to healthy subjects resulted in a small (11%)
`increase in mephenytoin biotransformation, suggestive of a minor induction of CYP2C19. In
`vitro studies have shown that vilazodone is a moderate inhibitor of CYP2C19 and CYP2D6.
`
`Thorough QT Study: Treatment with vilazodone does not prolong the QTc interval. The effect
`of vilazodone (20, 40, 60, and 80 mg) on the QTc interval was evaluated in a randomized,
`placebo-, and active-controlled (moxifloxacin 400 mg), parallel-group, thorough QTc study in
`157 healthy subjects. The study demonstrated an ability to detect small effects. The upper bound
`of the 90% confidence interval for the largest placebo-adjusted, baseline-corrected QTc interval
`was below 10 msec for vilazodone, based on the individual correction method (QTcI). This is
`below the threshold for clinical concern. However, it is unknown whether 80 mg is adequate to
`represent a high clinical exposure condition.
`
`
`5.0
`
`5.1
`
`Phase II Dose Finding Trials
`
`Five dose-finding trials were conducted as part of the Phase II program for vilazodone. Three of
`these trials were of flexible dose design (244, 245, and 247) and two were of fixed dose design
`(246 and 248). These were all double-blind, randomized, placebo-controlled, 8-week, parallel
`group trials in outpatients meeting DSM-IV criteria for MDD. The HAM-D-17 was the primary
`efficacy assessment and change from baseline to endpoint on the HAM-D-17 total score was the
`primary endpoint in all five trials. Three of these trials included an active comparator for assay
`sensitivity. Dosing in these trials was as follows:
`
`CLINICAL DATA
`
`Efficacy Data
`
`
`Reference ID: 2892948
`
`7
`
`

`

`
`
`
`Trial
`Number
`244
`245
`246
`247
`248
`
`Table 2: Dosing in Phase II Dose-Finding Trials
`Vilazodone Dosing (mg/day)
`Flexible-Dose Trials
`Fixed-Dose Trials
`
`5
`10
`20
`20-100
`
`
`
`10-20; 40-60-; 80-100
`
`
`
`---
`
`X
`X
`5-20
`
`
`
`---
`X
`X
`X
`
`Active
`Comparators
`(mg/day)
`Fluoxetine 20
`Fluoxetine 20
`Citalopram 20
`---
`---
`
`
`In none of these five trials was there a significant finding of superiority for any of the vilazodone
`trial arms vs. placebo on the primary endpoint, nor were any of the three active comparator arms
`shown to be superior to placebo on the primary endpoint. Thus, the three trials that included
`active control arms were judged to be “failed” trials, and the two vilazodone alone trials were
`judged to be “negative” trials. Although these trials were either failed or negative on the primary
`endpoint, the two fixed dose trials did reveal findings on the MADRS, a secondary endpoint, that
`were at least suggestive of efficacy at the 20 mg/day dose for vilazodone. Both trials included
`substantial sample sizes, i.e., about 120 per arm for study 246 and about 130 per arm for study
`248. The efficacy data for the MADRS total score are summarized in Table 3:
`
`
`Table 3: Summary of Efficacy Results for Fixed Dose Phase 2 Trials on MADRS
`(Difference between vilazodone and placebo in least square mean change from
`baseline using an LOCF approach)
`Dose
`Difference (Vilazodone-Pbo)
`10
`-2.3
`20
`-2.8
`5
`-0.4
`10
`-1.9
`20
`-2.5
`
`Trial
`246
`
`248
`
`
`
`P-Value
`P=0.123
`P=0.059
`P=0.725
`P=0.158
`P=0.062
`
`
`The phase 2 trials also revealed a dose response for adverse events within this broad vilazodone
`dose range, with increasingly poor tolerability as the dose was pushed above 40 mg/day. Figure
`4 (from the OCP review) depicts the dose-MADRS and AE relationships in these phase 2 trials.
`The depression symptoms decrease with increasing dose. Doses beyond 20 mg do not offer
`additional reduction in MADRS. However, certain AEs, e.g., abnormal dreams and nausea,
`increase with doses between 5 mg – 80 mg.
`
`
`Reference ID: 2892948
`
`8
`
`

`

`Figure 4. Dose response findings for efficacy and selected adverse events in phase 2 trials
`for vilazodone.
`
`
`
`
`
`
`Phase III Efficacy Trials
`
`Two nearly identical efficacy trials were conducted in the Phase 3 program for vilazodone
`(Studies 04 and 07). These were multicenter (all US sites), randomized, double-blind, parallel
`group, placebo-controlled, short-term (8-week) trials of vilazodone in adult patients (ages 18 to
`70) meeting DSM-IV-TR criteria for MDD, single episode or recurrent. Patients were required
`to have at screening and baseline visits a HAM-D score of > 22 on the first 17 items of the 21
`item HAM-D, and a HAM-D item 1 (depressed mood) score of > 2. Both studies included only
`a single fixed dose of 40 mg/day. Also in Study 04, patients who could not tolerate the 40 mg
`dose could be maintained on 20 mg/day. Randomization was 1:1 vilazodone 40 mg/day vs.
`placebo in both studies. Vilazodone was initiated at 10 mg/day for 7 days, then increased to 20
`mg/day for 7 days, and maintained at 40 mg/day for weeks 3 through 8. Vilazodone was taken
`with food in both studies 04 and 07. The 40 mg/day dose was selected based in part on the
`finding in the phase 2 trials that 20 mg/day appeared to be on the margin of an effective dose
`range and the PET study showing that 40 mg/day is required to have any significant occupancy
`at the 5HT1A receptor. The MADRS was the primary efficacy assessment, and was conducted
`at baseline and weeks 1, 2, 4, 6, and 8. The primary endpoint was change from baseline to week
`8 on the MADRS total score. The primary analysis was analysis of covariance (last observation
`carried forward) for a modified intent-to-treat (ITT) population. These were the patients who
`took at least one dose of their assigned treatment and who had baseline and at least one follow-
`up efficacy assessment. CGI-I and CGI-S were included among several secondary endpoints.
`
`
`
`Reference ID: 2892948
`
`9
`
`

`

`CLDA-07-DP-02 (Study 7): This multicenter (15 sites) US study compared vilazodone 40
`mg/day vs placebo (1:1). There were about 230 patients per group (ITT). There were about
`19% dropouts (roughly the same for both groups). Placebo had a higher % dropout for lack of
`efficacy and a lower % dropout for AEs, compared to vilazodone. The mean age for the sample
`was 42 years and the female:male ratio was 56%:44%. The mean baseline MADRS total score
`was 32, and the least square mean changes from baseline for MADRS to week 8 were: -10.8
`(pbo) and -13.3 (vilazodone). The difference between groups in change from baseline was -2.5
`(SE=0.96; 95% CI=-4.4, -0.6; p=0.009). A mixed-effects model for repeated measures
`(MMRM) analysis also favored vilazodone vs. placebo, as did analyses of CGI-S and CGI-I.
`
`GNSC-04-DP-02 (Study 4): This multicenter (18 sites) US study compared vilazodone 40
`mg/day vs placebo (1:1). Patients who could not tolerate the 40 mg dose could be reduced to 20
`mg/day. There were about 204 patients per group (ITT). There were about 25% dropouts
`(roughly the same for both groups). Placebo had a higher % dropout for lack of efficacy and a
`lower % dropout for AEs, compared to vilazodone. The mean age for the sample was 40 years
`and the female:male ratio was 63%:37%. The mean baseline MADRS total score was 31, and
`the least square mean changes from baseline for MADRS to week 8 were: -9.7 (pbo) and -12.9
`(vilazodone). The difference between groups in change from baseline was -3.2 (SE=0.99; 95%
`CI=-5.1, -1.2; p=0.001). A mixed-effects model for repeated measures (MMRM) analysis also
`favored vilazodone vs. placebo, as did analyses of CGI-S and CGI-I.
`
`Forty-one patients were maintained on the middle dose due to reasons of intolerability, i.e., 20
`mg in the vilazodone group and 2 placebo pills for the placebo group. There were 28 such
`patients in the vilazodone group and 13 in the placebo group. Thirteen of these patients
`completed the study, i.e., 6 on vilazodone and 5 on placebo. The drug minus placebo difference
`in least square mean change from baseline on the MADRS total score was -4.3 (95% CI -
`11.6,2.9; P=0.23).
`
`Onset of Effect: FDA conducted mixed model repeated measures (MMRM) analyses for trials 7
`and 4 as sensitivity analyses. These analyses provide results by treatment visit over the course of
`these 8-week trials, as illustrated in Table 4:
`
`Table 4: Time of Onset of Statistically Significant Treatment Effect in Trials 7 and 4
`(Difference between drug and placebo in change from baseline in MADRS total
`score, based on MMRM Analyses)
`
`Trial 7
`Visit Week
`Difference
`1
`-0.4
`2
`-1.0
`4
`-1.6
`6
`-2.3
`8
`-2.9
`
`Whether or not the nominally statistically significant vilazodone-placebo treatment differences
`observed at these earlier time points (week 4 for Trial 7 and week 1 for Trial 4) represent
`clinically relevant treatment effects is unknown.
`
`P-Value
`P=0.347
`P=0.087
`P=0.050
`P=0.017
`P=0.006
`
`Trial 4
`Difference
`-1.7
`-1.7
`-2.9
`-4.1
`-3.6
`
`P-Value
`P=0.0001
`P=0.0063
`P=0.0005
`P<0.0001
`P=0.0007
`
`
`Reference ID: 2892948
`
`10
`
`

`

`
`Subgroup Analyses: Subgroup analyses for these 4 studies based on gender, age, and race
`generally showed consistency in the results across these subgroups.
`
`DSI: DSI inspected 4 clinical sites, and found the data generated for this program to be
`acceptable.
`
`-Efficacy Conclusions: I agree with Drs. Dinh, Lindberg, and Levin that the sponsor has
`demonstrated efficacy for vilazodone in the treatment of MDD at a dose of 40 mg/day. The
`sponsor has not conducted a maintenance study in MDD, but they have committed to conducting
`such a study. In addition, they have committed to conducting a pediatric MDD study (ages 7-
`17), and a study at 20 mg in adults.
`
`Genetic Data
`
`Genotype data were available for assessment of the relationships between genetic variations in
`CYP2C19, CYP2D6, and ACE with vilazodone response in the two phase 3 studies. There were
`no meaningful associations with MADRS by responder analyses or change at 8 weeks. There
`were also no robust associations with discontinuation rates or failure to reach target dose that
`could explain the similar responses in CYP2C19 UMs and PMs. Genotype data were not
`available for all pivotal studies in which DNA was collected. Given the complicated vilazodone
`metabolic pathway, it is uncertain whether genetic variation on pharmacokinetic-related genes
`would likely result in clinically meaningful differences in either response or adverse events.
`
`5.2
`
`The development program for vilazodone in MDD included data from 24 phase 1 studies, 5
`phase 2 studies, and 3 phase 3 studies, all in adults. These 32 studies included a total of 2898
`adult subjects exposed to one or more doses of vilazodone. The 24 phase 1 program involved
`721 subjects exposed to vilazodone doses ranging from 1 to 80 mg in single and repeat dose
`studies. FDA’s safety review focused primarily on the 8 phase 2-3 studies, involving 2177
`patients exposed to vilazodone. Seven of these were placebo-controlled, 8-week studies that
`included 1578 patients with MDD exposed to vilazodone. The eighth study was an open label
`study involving 599 patients exposed to vilazodone for up to one year. Overall, the vilazodone
`exposure was 552 subject-years. The phase 2-3 studies included only 37 patients > 65 years old
`and only 272 in the 55 to 64 year old age range. There were no pediatric patients exposed to
`vilazodone in this program.
`
`Deaths: There were 3 deaths in the development program, including a homicide in a phase 1
`study, and 2 suicides in later studies. Neither suicide occurred in a patient exposed to
`vilazodone.
`
`Nonfatal Serious Adverse Events (SAEs): Overall, 81 patients experienced a nonfatal SAE,
`including 5 in phase 1 studies and 76 in phase 2-3 studies.
`-Phase 1: Of these 5 patients with SAEs, 4 occurred in vilazodone patients, and one in an active
`control patient. Of the 4 vilazodone events, 2 were psychiatric (1 occurred 5 days after stopping
`
`Safety Data
`
`
`Reference ID: 2892948
`
`11
`
`

`

`drug), 1 was a fracture, and 1 was a seizure, possibly related to vasovagal syncope occurring
`during a blood draw. The other psychiatric event may have been drug-related, however, it was
`not clear that this patient was actually a “healthy volunteer”. Overall, the proportion of SAEs
`was similarly low in vilazodone [4/721(0.6%)] and placebo patients [0/132(0%)].
`-Phase 2: There were 37 patients with SAEs in the phase 2 program, including 21 vilazodone
`patients, 13 placebo patients, and 3 active control patients. Of the 21 events in vilazodone
`patients, 6 were psychiatric, 5 were “pregnancy”, 1 was “medication missing”, and the remaining
`9 were isolated events that are all common background events: diarrhea; hypertension; tremor,
`migraine; TIA; fracture; atrial fibrillation; cervicitis; pneumonia. Diarrhea and tremor are
`established adverse reactions with vilazodone; the remaining 7 events were not, in view,
`plausibly drug-related. Overall, the proportion of SAEs was similar in vilazodone
`[21/1137(1.8%)] and placebo patients [13/562(2.3%)]. There was a similar broad scattering of
`events among vilazodone and placebo patients, with no pattern implicating vilazodone as a cause
`of any other than the two already established events.
`-Phase 3 Controlled Trials: There were 16 patients with SAEs in the two controlled trials in the
`phase 3 program, including 9 in vilazodone patients and 7 in placebo patients. Of the 9 events in
`vilazodone patients, 2 were psychiatric and 7 were isolated common background events: prostate
`cancer; concussion; lymphoma; CAD; pleural effusion; chest pain; and cholecyctitisis, none of
`which were plausibly drug-related, in my view. Overall, the proportion of SAEs was similar in
`vilazodone [9/436(2.0%)] and placebo patients [7/433(1.6%)]. There was a similar broad
`scattering of events among vilazodone and placebo patients, with no pattern implicating
`vilazodone as a cause of any of the SAEs in these 2 trials.
`-Phase 3 Open Label Trial: There were 23 patients with SAEs in the open label phase 3 trial.
`Three of these were psychiatric events, and the remaining covered the spectrum of non-
`psychiatric events. With the exception of 2 instances of pneumonia, these were isolated,
`common background events, with no pattern implicating vilazodone as a cause for any of these.
`
`These events are difficult to assess for causality. Overall, however, for the phase 2-3 studies, the
`proportions of patients experiencing SAEs appeared to be similar for vilazodone and placebo
`patients. Many of the SAEs were psychiatric events that likely represented worsening of the
`underlying condition being treated, and are expected in any psychiatric drug development
`program. The others were common background events, with no pattern of findings suggesting
`any particular event was more common in patients exposed to vilazodone than those exposed to
`placebo.
`
`Dropouts for Adverse Events: In the placebo controlled trials with vilazodone, there was no
`single adverse event leading to discontinuation in > 1% of patients. Approximately 7% of
`vilazodone-exposed patients discontinued due to an adverse event compared to 3% of placebo-
`exposed patients. The most common adverse events leading to discontinuation were
`gastrointestinal, psychiatric, and neurological.
`
`Predicted Adverse Events: Given that vilazodone is an SSRI, certain adverse events might be
`predicted, and several of these were observed:
`-Serotonin Syndrome: The database was searched for terms suggestive of possible serotonin
`toxicity. Two subjects were identified with such events, including one patient in the long-term
`
`
`Reference ID: 2892948
`
`12
`
`

`

`open label phase 3 study. This patient took an overdose of approximately 240 mg of vilazodone.
`The other patient had received a vilazodone dose of 80 mg in a phase 2 study.
`-Mania/Hypomania: A search of the database discovered 6 patients with likely treatment-
`emergent mania or hypomania, including 5 taking vilazodone and 1 taking placebo.
`-Bleeding: Although bleeding events were identified in the vilazodone database, the proportions
`of patients experiencing such events were similar in vilazodone and placebo groups, i.e.,
`approximately 3% for each.
`
`Common Adverse Events: The following table (Table 5), derived from a pooling of the two
`phase 3 trials, includes adverse events occurring at a rate of at least 5% on vilazodone and
`having a rate at least twice the placebo rate, i.e., those FDA considers common and drug-related.
`
`
`Table 5: Common and Drug-Related Adverse Events
`Adverse Event Term
`Vilazodone 40 mg
`Placebo
`Diarrhea
`28%
`9%
`Nausea
`23%
`5%
`Vomiting
`5%
`1%
`Insomnia
`6%
`2%
`
`
`Additional adverse events identified from this pooling as having a rate of at least 2% and at least
`twice the placebo rate included the following: gastroenteritis, paresthesia, tremor, abnormal
`dreams, restlessness, libido decreased, orgasm abnormal, delayed ejaculation, erectile
`dysfunction, feeling jittery, palpitations, and increased appetite. There is strong evidence for
`dose-relatedness for many of the common adverse events, with doses above 40 mg being poorly
`tolerated for most patients.
`
`Labs, Vital Signs, Weight, ECGs: Vilazodone was not associated with any clear finding of drug-
`related changes in laboratory parameters, vital signs, or weight in the placebo-controlled trials.
`Laboratory testing included routine serum chemistries, thyroid testing, routine hematology
`testing, and urinalysis. Vital signs included blood pressure, pulse, respiratory rate, and ECGs. A
`thorough QT study with vilazodone did not reveal QT prolongation or any other important
`changes in ECG parameters.
`
`Ophthalmological Findings: Concern about ophthalmological toxicity with vilazodone was
`initially based on a finding of transient corneal opacities in a dog study. Longer-term animal
`studies did not, however, confirm the early signal for corneal pathology. Thus, there was no
`confirmed signal for corneal pathology in animal studies, and never even a hint of cataract
`formation in animals. Nevertheless, based on the early animal findings, certain human studies
`included various ophthalmologic evaluations, i.e., slit lamp exams, dilated fundoscopy, tests of
`visual acuity, intraocular pressure, and Schirmer’s test for decreased lacrimation. These tests
`were done only in the 5 phase 2 studies and some in the uncontrolled, longer-term phase 3 study.
`
`The only clear finding from the extensive exams in the phase 2 studies was a slight decrease in
`tear production based on results of Schirmer testing, and this was the likely explanation for at
`least some of the reports of blurred vision and the few reports of corneal abnormalities. The
`small number of reports of retinal or vitreous findings in these phase 2 studies were not
`
`
`Reference ID: 2892948
`
`13
`
`

`

`suggestive of any drug related problem in these areas. A question had been raised in a 5-7-01
`review by our ophthalmological consultant, Dr. Wiley Chambers, regarding a possible signal for
`treatment-emergent cataract formation from 2 of the phase 2 studies, however, in retrospect, Dr.
`Chambers has now acknowledged that the actual increase in cataracts was observed in the active
`control arm, while the vilazodone arms had cataract findings similar to the placebo arm.
`
`Eye-related adverse events for the year-long open label study included dry eye (4.7%) and
`blurred vision (4.0%), and these were like related to the dry eye problem detected in phase 2
`studies. Of interest, these problems did not emerge as drug-related in the 2 phase 3 studies.
`There were other much less common ophthalmological adverse events reported in this open label
`study, but with a pattern and frequency consistent with the background rate observed in this
`population. The only other finding of interest in this open label study was a 13% rate of cataract
`progression over 1 year, i.e., detectable worsening of existing cataracts. Dr. Chambers
`considered this progression rate somewhat high, but it is hard to interpret in the absence of a
`control group. On the other hand, we did find a reference for a study that used an LOCSII
`approach at 6-month intervals to assess the rate of progression for patients with cataracts
`(LOCSII was also used in the 52-week vilazodone study). After 6 months, worsening could be
`detected