CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`
`APPLICATION NUMBER:
`
`022567Orig1s000
`
`OFFICE DIRECTOR MEMO
`
`
`
`
`

`

`Deputy Office Director Decisional Memo
`
`1/21/2010
`Ellis F. Unger, M.D., Deputy Director, ODE-I
`Deputy Office Director Decisional Memo
`22-567
`PGx Health, LLC
`March 22, 2010
`January 22, 2011
`Viibryd
`Vilazodone Hydrochloride
`Oral tablets: 10 mg, 20 mg, and 40 mg
`… for the treatment of major depressive disorder (MDD)
`Approval
`
`
`Date
`From
`Subject
`NDA#
`Applicant Name
`Date of Submission
`PDUFA Goal Date
`Proprietary Name /
`Established (USAN) Name
`Dosage Forms / Strength
`Indication
`Action:
`
`Material Reviewed/Consulted
`Action Package, including:
`Project Manager
`Medical Officer Clinical Review
`Clinical Pharmacology Review
`
`William Bender
`Cheri Lindberg/Robert Levin (team leader)
`Bei Yu/Huixia Zhang/Jee Eun Lee/Jogarao
`Gobburu/Venkatesh Bhattaram/Yaning
`Wang/Issam Zineh
`Phillip Dinh/Peiling Yang (team leader)
`Violetta Klimek/ Linda Fossom (team leader)
`Pei-I Chu/ Chhagan Tele (team leader)
`Mohamed Nagem/Karl Lin (team leader)
`
`Pei-I Chu/ Chhagan Tele (team leader)
`Anthony Orencia/Purohit-Sheth, Tejashri
`Loretta Holmes/Kristina Tolliver (team leader)
`
`Shawna Hutchinson – REMS
`Robin Duer - Med Guide
`Jessica Cleck Derenick
`
`Robert Levin/Mitchell Mathis
`Loretta Holmes
`Thomas Laughren
`
`Statistical Review
`Pharmacology Toxicology
`Chemistry Manufacturing and Controls
`Statistical Review and Evaluation of
`Carcinogenicity Study
`Environmental Assessment
`Division of Scientific Investigations
`Division of Medication Error Prevention and
`Analysis, Office of Surveillance and Epidemiology
`Division of Risk Management,
`Office of Surveillance and Epidemiology
`Division of Drug Marketing, Advertising and
`Communications (DDMAC)
`Cross-Discipline Team Leader
`Proprietary Name Review
`Director, Division of Psychiatry Products
`
`Action:
`
`The Division of Psychiatry Products is recommending approval of vilazodone hydrochloride, 10-,
`20-, and 40-mg Tablets for oral administration for the treatment of major depressive disorder
`(MDD). I concur with their recommendation for approval.
`
`
`
`Reference ID: 2895106
`
`Page 1
`
`

`

`Introduction:
`
`Vilazodone is a small molecule, both a selective serotonin reuptake inhibitor (SSRI) and a
`partial 5-HT1A receptor agonist, although the clinical significance of the latter is unknown.
`Vilazodone is not marketed anywhere.
`
`Regulatory Background:
`
`The IND for vilazodone for this indication was submitted on 11/21/1997 by Lipha
`Pharmaceuticals. Sponsorship of the IND was transferred numerous times, as summarized in
`Dr. Levin’s CDTL review. The applicant had originally hoped to use genetic markers to direct
`clinical decision-making; however, those plans were ultimately abandoned. A pre-NDA meeting
`was scheduled for June, 2009; however, the applicant found the Division’s preliminary
`comments sufficient to address their questions and the meeting was cancelled.
`
`Chemistry Manufacturing and Controls (CMC):
`
`Pursuant to their initial review, the CMC team sent an information request (IR) letter to the
`applicant on 10/15/2010, and the applicant’s responses were deemed adequate. The ONDQA
`Biopharmaceutics review the found the proposed dissolution methodology and specifications to
`be acceptable. The Environmental Assessment review found no pending issues. Accordingly,
`the NDA was recommended for approval from a CMC perspective.
`
`Pharmacology/Toxicology:
`
`The review found the application approvable. Vilazodone binds with high affinity to the
`serotonin reuptake site (Ki = 0.1 nM), but not to the dopamine or norepinephrine reuptake sites.
`Vilazodone inhibits reuptake of serotonin (IC50 = 1.6 nM) and binds to 5-HT1A receptors with an
`IC50 of 2.1 nM, where it functions as a partial agonist.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` Moreover, their proposed
`
`Reference ID: 2895106
`
`Page 2
`
` Serotonergic mechanisms in the
`central nervous system (CNS) are complex. Experimentally, vilazodone has been observed to
`exhibit both agonism and antagonism, depending on the experimental model and region of the
`brain studied. Moreover, 5-HT1A receptors are present at both presynaptic and postsynaptic
`nerve terminals, and their various interactions are not fully understood. In short, the net effect of
`5-HT1A partial agonism on serotonergic transmission in the CNS has not been well-
`characterized, and the clinical significance of those effects, if any, is certainly unknown.
`
`
`
`
`
`
`
`
`
`
`
`(b) (4)
`
`(b) (4)
`
`(b) (4)
`
`

`

` Through much
`proprietary name, “Viibryd,”
`discussion and negotiation, the review team was able to reach agreement with the applicant on
`a description of vilazodone’s mechanism of action (section 12.1 of the label):
`
`
`“The mechanism of the antidepressant effect of vilazodone is not fully understood but is
`thought to be related to its enhancement of serotonergic activity in the CNS through
`selective inhibition of serotonin reuptake. Vilazodone is also a partial agonist at
`serotonergic 5-HT1A receptors; however, the net result of this action on serotonergic
`transmission and its role in vilazodone’s antidepressant effect are unknown.”
`
`
`
`
` It is notable that the applicant has made
`no effort to show a benefit in comparison to other antidepressants; no comparators were
`included in the phase 3 program. Vilazodone has two major human metabolites, M10 and M17,
`each circulating at greater than 10% of total drug-related exposure. Neither is thought to have
`important serotonergic activity. Both have been assessed for toxicity; however, it is unclear if
`M17 has been adequately assessed for embryo-fetal toxicity because it was not found to be
`present in the plasma of either rats or rabbits. The applicant has agreed to explore this issue in
`a reproductive toxicity study, post-approval, in which M17 is administered by a route that will
`produce systemic exposure greater than or equal to the exposure in humans at the maximum
`recommended human dose (MRHD). Alternatively, they could show that the original rabbit
`study was adequate to assess the embryo-fetal toxicity of M17 by demonstrating that the
`rabbit’s systemic exposure to M17 in that study was greater than or equal to that in humans at
`the MRHD.
`
`Two-year carcinogenicity studies were conducted in B6C3F1 mice and Wistar rats, given oral
`vilazodone at approximately 16.5 and 36 times the MRHD, respectively. The studies were
`deemed to be acceptable. In mice, the incidence of hepatocellular carcinomas was increased in
`males at 16.5 times, but not 5.5 times, the MRHD. Mammary gland adenocarcinomas were
`increased in females at 5.5 and 16.5 times the MRHD (associated with increased prolactin
`levels – known to cause mammary tumors in rodents), but not at 1.8 times the MRHD. In rats,
`there were no biologically relevant drug-related increases in incidences of neoplasms at doses
`up to 36 times the MRHD.
`
`
`Results of mutagenicity assays were mixed: vilazodone was clastogenic in vitro in assays for
`chromosomal aberrations using V79 CHO cells in the presence and in absence of S9 metabolic
`activation, and using human lymphocytes in the presence of S9 activation. Vilazodone was
`negative for mutagenicity in the Ames test and in the hypoxanthine-guanine phosphoribosyl
`transferase (HPRT) assay. It was also negative in several in vivo studies that included: 1) a
`chromosomal aberration assay in the rat bone marrow cells; 2) a micronucleus test in rats; and
`3) an unscheduled DNA synthesis (UDS) assay in rat hepatocytes.
`
`Treatment of rats with vilazodone at 30 times, but not 6 times, the MRHD caused impairment of
`male fertility; there was no effect on female fertility.
`
`Vilazodone caused developmental toxicity in rats (reduced fetal weight and delayed bone
`ossification) but was not teratogenic in either rats or rabbits.
`
`Relevant to gastrointestinal adverse reactions observed in the phase 2 and 3 clinical trials, there
`were no notable effects on gastrointestinal transport or gastric emptying in rodents in the safety
`pharmacology studies.
`
`Reference ID: 2895106
`
`Page 3
`
`(b) (4)
`
`(b) (4)
`
`

`

`
`Site Inspections:
`
`Four U.S. investigators (2 in each of the phase 3 trials) were inspected in support of this NDA;
`the applicant was inspected as well. Because larger sites were selected for inspection, the
`Division of Scientific Investigations had access to records of 311 subjects at 6 centers during
`their inspections, or approximately 35% of subjects in the phase 3 trials.
`
`Minor regulatory deficiencies were found in one of the studies, but they were isolated and
`thought to have minimal impact on either data integrity or protection of human subjects. Overall,
`the data were deemed reliable for the proposed indication, with general adherence to Good
`Clinical Practices (GCP) regulations governing the conduct of clinical investigations.
`
`Pharmacokinetics:
`
`Vilazodone exhibits dose-proportional pharmacokinetics over a dose range from 5 to 80 mg.
`Administration with a high-fat or light meal increases oral bioavailability, and when administered
`with food, vilazodone’s absolute bioavailability is ~72%. The applicant proposes that vilazodone
`be taken with food, as it was in the phase 3 program, and the Division agrees with this
`recommendation. The median Tmax is 4-5 hours, and terminal half-life is ~25 hours.
`
`Vilazodone is widely distributed, with a volume of distribution of 600 L after a 5 mg infusion, and
`the drug is highly protein-bound (96-99%).
`
`Vilazodone’s accumulation is predictable from single-dose data (accumulation factor of about
`1.8, and consistent across different doses), and steady state is achieved in ~3 days. The mean
`steady state Cmax after daily 40 mg dosing under fed conditions is ~160 ng/mL.
`
`Vilazodone is extensively metabolized through CYP and non-CYP pathways, with 1% and 2% of
`the unchanged drug recovered in urine and feces, respectively. Among the CYP pathways,
`CYP3A4/5 is principally responsible for vilazodone’s metabolism, with only minor contributions
`from CYP2C19 and CYP2D6. In vitro studies show that vilazodone is unlikely to inhibit or
`induce the metabolism of other CYPs (except for CYP2C8). Strong CYP3A4 inhibitors can
`reduce vilazodone’s metabolism and increase exposure modestly (coadministration with
`ketoconazole increases the AUC and Cmax by 50%). The label includes a recommendation to
`reduce the vilazodone dose from 40 to 20 mg daily when administered with strong CYP3A4
`inhibitors. Theoretically, CYP3A4 inducers (e.g., carbamazepine) might decrease vilazodone
`exposure, although this was not studied. The applicant has agreed to conduct a drug-drug
`interaction trial of vilazodone using carbamazepine in healthy subjects as a postmarketing
`commitment.
`
`Vilazodone had minimal effects on other drugs, except that coadministration of vilazodone with
`a CYP2C8 substrate can lead to an increase in the concentration of the other drug.
`
`Thorough QT Study:
`
`As described by others, vilazodone was tested in a thorough QT study at doses up to 80 mg.
`The study demonstrated appropriate assay sensitivity, and the baseline-corrected QTc interval
`was <10 msec for vilazodone, below the threshold of clinical concern.
`
`Phase 2 Dose-Finding Trials:
`
`Reference ID: 2895106
`
`Page 4
`
`

`

`
`The Phase 2 program included 5 dose-finding studies. Three incorporated flexible-dose designs
`(244, 245, and 247) and 2 were fixed-dose designs (246 and 248). The studies were quite
`similar: all were 8-week, randomized, double-blind, placebo-controlled, parallel group studies in
`adult outpatients meeting DSM-IV criteria for MDD. Each study enrolled between 86 and 140
`subjects per treatment group. The Hamilton Rating Scale for Depression (HAM-D) was the 1°
`efficacy endpoint, assessed as change from baseline to Week 8 on the sum the scores for the
`first 17 items (HAM-D-17). Studies 244, 245, and 246 included an active comparator to assess
`assay sensitivity. The dosing paradigms and active comparators are shown in Table 1:
`
`Table 1: Phase 2 Dose-Finding Studies
`
`
`Trial
`number
`
`Vilazodone Dosing
`Fixed-dose studies
`Flexible-dose studies
`5 mg/d
`10 mg/d 20 mg/d
`
`Active comparators
`(mg/day)
`
`244
`245
`246
`247
`248
`
`20-100 mg/d
`10-20; 40-60; 80-100 mg/d
`---
`5-20 mg/d
`---
`
`X
`
`X
`
`X
`
`X
`
`X
`
`fluoxetine 20
`fluoxetine 20
`citalopram 20
`---
`---
`
`
`None of the 5 studies showed a statistically significant treatment effect of vilazodone on its 1°
`endpoint; in fact, none were even close. None of the 3 active comparator studies showed the
`comparator to be statistically distinguishable from placebo. It could be concluded, therefore,
`that the 2 studies lacking an active control group were “negative” trials
`
`
`
`
`
`
`
`
`
`
`
`
`
`Although these studies were unsuccessful on their 1° endpoints, the 2 fixed-dose trials (246 and
`248) showed dose-responses on a 2° endpoint, the Montgomery-Asberg Depression Rating
`Scale (MADRS), suggestive of a treatment effect for the 20 mg/day vilazodone groups (the
`nominal unadjusted p-values were 0.06 in both studies). The 2° endpoint data for the MADRS
`total score are summarized in Table 2:
`
`Table 2: Efficacy Results for Fixed-Dose Phase 2 Trials on the 2° MADRS Endpoint
`
`
`Reference ID: 2895106
`
`Page 5
`
`(b) ( )
`
`(b) (4)
`
`

`

`study
`
`dose
`(mg/d)
`
`ℵMADRS (vilazodone
`minus placebo)
`
`p-value
`
`246
`
`248
`
`10
`20
`5
`10
`20
`
`-2.3
`-2.8
`-0.4
`-1.9
`-2.5
`
`0.12
`0.06
`0.72
`0.16
`0.06
`
`
`
`
`
`Evidence of Effectiveness:
`
`The applicant submitted two phase 3 trials to establish the evidence of effectiveness for
`vilazodone for the sought indication “treatment of MDD.” Results of one of the two phase 3
`trials has been published: Rickels K, et al. J Clin Psychiatry 2009;70:326-33.
`
`These were identified by the applicant as studies GNSC-04-DP-02 and CLDA-07-DP-02, and
`are referred to as Studies 4 and 7, respectively, in this memorandum.
`
`Design: Both were multicenter, randomized, double-blind, placebo-controlled studies of
`vilazodone in adults meeting DSM-IV-TR criteria for MDD, with a single or recurrent episode.
`Both were short-term studies, 8 weeks in length – the same length as the failed phase 2 studies.
`
`Both phase 3 studies randomized subjects 1:1 to vilazodone or placebo, and the vilazodone
`dose was ultimately fixed at 40 mg/d; however, the dose was gradually “ramped” to decrease
`side effects: 10 mg/d during Week 1, 20 mg/d in Week 2, and finally 40 mg/d during the last 6
`weeks (Weeks 3 to 8). In Study 4 (only), patients who could not tolerate the 40 mg daily dose
`could be maintained on 20 mg/d. Vilazodone was to be taken with food to enhance
`bioavailability.
`
`Dose Selection: Selection of dose was based on the phase 2 findings, wherein the 20 mg/d
`dose trended towards showing a treatment effect, as well as a positron emission tomography
`(PET) study showing that 40 mg/day is required to achieve significant occupancy of 5-HT1A
`receptors. Adverse events such as dizziness and abnormal dreams were dose-dependent,
`leading to higher drop-outs at doses of 60 mg/d. Of note, a 40 mg/d fixed-dose had not been
`evaluated in phase 2.
`
`For both studies, subjects were required to have a total HAM-D-17 of >22 and a HAM-D
`depressed mood score (item 1 of HAM-D) of >2 at both screening and baseline visits.
`
`Endpoints: Not surprisingly, in light of the phase 2 results, the 1° efficacy endpoint was the
`MADRS total score (and not the more commonly used HAM-D), assessed as change from
`Baseline to Week 8. The analytic approach was an analysis of covariance (last observation
`carried forward, LOCF), and the intent-to-treat (ITT) population was the population of interest,
`defined as subjects who received ≥1 dose of their assigned treatment and who had ≥1 post-
`baseline efficacy assessment. There were numerous 2° efficacy endpoints, including: MADRS
`response, MADRS remission, Δ HAM-A total score, Δ HAM-D, HAM-D response, HAM-D
`
`Reference ID: 2895106
`
`Page 6
`
`

`

`remission, HAM-A total score, Δ Clinical Global Impressions-Severity (CGI-S), Clinical Global
`Impressions-Improvement (CGI-I), and CGI-I response. Because there was no prospective plan
`to control Type-I error for the plethora of 2° endpoints, they should be considered exploratory in
`nature.
`
`Results – Study 4:
`Study 4 was conducted from February, 2006 to May, 2007. The statistical analysis plan was
`finalized 2/27/2007, with amendments 5/17/2007 and 7/13/2007.
`
`Study 4 was conducted at 18 sites in the US. A total of 561 patients were screened to enroll
`205 subjects per treatment group, with 1/6 of screened patients failing to meet entrance criteria.
`Approximately 25% of subjects discontinued in both treatment groups. Twice as many subjects
`discontinued because of an adverse event in the vilazodone group relative to the placebo group,
`whereas approximately half as many discontinued for lack of efficacy, and one-third as many
`withdrew consent. The ITT population included 97% of the enrolled subjects in both groups.
`
`Demographic and baseline disease-related variables were reasonably matched between
`treatment groups. Mean age was 40 years; subjects were 63% female and 14% black. The
`mean baseline MADRS total score was 31. The least square mean changes for MADRS
`(Baseline to Week 8) were -9.7 in the placebo group and -12.9 in the vilazodone group. The
`difference between groups was -3.2 (standard error [SE] = 0.99; 95% CI = -5.1, -1.2; p = 0.001).
`A mixed-effects model for repeated measures (MMRM) analysis supported the 1° efficacy
`analysis.
`
`Figure 1 shows the distribution of changes in MADRS for the 2 treatment groups (a more readily
`interpretable variation on the cumulative distribution function curves, generated by Dr. Dinh for
`this study, and typically shown for Alzheimer’s Disease drugs), with each “bin” representing a 5-
`point ΔMADRS. The figure was generated from the data in the SAS transport file
`0000\m5\datasets\gnsc-04-dp-02\analysis\a-madr.xpt, using LOCF to impute missing data.
`
`
`Reference ID: 2895106
`
`Page 7
`
`

`

`Vilazodone
`
`Placebo
`
`+15 to +19
`+10 to +14
`+5 to +9
`
`4
`
`0 to +
`
`-5 to -1
`
`-10 to -6
`-15 to -11
`-20 to -16
`-25 to -21
`-30 to -26
`-35 to -31
`-40 to -36
`
`30
`
`25
`
`20
`
`15
`
`10
`
`05
`
`percent of subjects
`
`better Change in MADRS worse
`
`The effect on ΔMADRS is modestly shifted to the left (“better”) in vilazodone-treated subjects.
`
`An analysis of covariance on the change from Baseline to Week 8 in the HAM-D-17 total score
`using the ITT population with missing values imputed by LOCF, the metric typically used to
`show evidence of efficacy for antidepressant drugs, supported the primary efficacy results.
`
`Of note, 41 subjects were maintained on the 20 mg/d dose (or equivalent in the placebo group)
`because of poor tolerability. There were 28 such subjects 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, and in this selected population, the difference in least squares mean change from
`baseline on the MADRS total score was -4.3 (95% CI -11.6, 2.9; p = 0.23). This trend suggests
`that the 20-mg daily dose might have efficacy in this patient population.
`
`Results – Study 7:
`
`Study 7 was conducted from March 31, 2008 to February 10, 2009. The statistical analysis plan
`was finalized February 3, 2009, with an amendment on May 13, 2009.
`
`Study 7 enrolled subjects at 15 US sites. A total of 659 patients were screened to enroll ~240
`subjects per treatment group, with 13% of screening failures related to inability to meet entrance
`criteria. Nineteen percent (19%) of subjects discontinued in each treatment group. The leading
`causes were “lost to follow-up” (7% in both groups), and withdrawal of consent (5% in both
`groups). The placebo group had a slightly higher rate of discontinuation for lack of efficacy (3%,
`versus 1% for vilazodone) and a lower rate for adverse events (2%, versus 5% for vilazodone).
`
`Demographic and baseline disease-related variables were similar in the two treatment groups.
`Mean age was 42 years: subjects were 56% female and 14% black. The mean baseline
`MADRS total score was 32. The least square mean changes for MADRS (Baseline to Week 8)
`were -10.8 in the placebo group and -13.3 in the vilazodone group, with a difference between
`
`Reference ID: 2895106
`
`Page 8
`
`

`

`groups of -2.5 (SE=0.96; 95% CI = -4.4, -0.6; p = 0.009). A MMRM analysis supported the 1°
`efficacy analysis.
`
`Figure 2 shows the distribution of changes in MADRS for the 2 treatment groups, with each “bin”
`representing a 5-point ΔMADRS. The figure was generated from the SAS transport file
`0000\m5\datasets\clda-07-dp-02\analysis\dmadrl.xpt, using LOCF for missing assessments.
`
`
`Vilazodone
`
`Placebo
`
`+5 to +9
`
`4
`
`0 to +
`
`-5 to -1
`
`-10 to -6
`
`1
`
`5 to -1
`
`-1
`
`1
`
`5 to -2
`
`-2
`
`6
`
`0 to -1
`
`-2
`
`6
`
`0 to -2
`
`-3
`
`6
`
`0 to -3
`
`-4
`
`1
`
`5 to -3
`
`-3
`
`35
`
`30
`
`25
`
`20
`
`15
`
`10
`
`05
`
`percent of subjects
`
`better Change in MADRS worse
`
`The effect on ΔMADRS is modestly shifted to the left (“better”) in vilazodone-treated subjects.
`The most apparent shift is from the minimally better category (-5 to -1) to the 3 categories
`ranging from -21 to -35 (approximately 12% of subjects).
`
`An analysis of covariance on the change from Baseline to Week 8 in the HAM-D-17 total score
`(ITT population; missing values imputed LOCF) supported the primary efficacy results.
`
`Onset of Effect: The FDA biostatistician conducted MMRM analyses for Trials 7 and 4 as
`sensitivity analyses, showing results as a function of time (Table 3):
`
`Table 3: ΔMADRS by Visit for Trials 7 and 4 (adapted from FDA Office of Biostatistics Review)
`
`
`
`
`Reference ID: 2895106
`
`Page 9
`
`
`
`

`

`The efficacy results as a function of time are somewhat surprising, in that a treatment effect is
`observed as early as 1 week after initiation of treatment (in Trial 4). Not only does this suggest
`a rapid onset of effect, but, perhaps even more surprising, the effect is observed at the starting
`vilazodone dose – only 10 mg/d – which is a mere quarter of the final, fixed 40-mg dose.
`
`The observation should be considered in perspective, however: the finding is present in only 1
`of 2 studies, and analyses of the 1° endpoint at multiple time points was not subjected to a
`rigorous, prospectively-designed statistical approach intended to consider multiplicity and Type-I
`error.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Subgroup Analyses: In Trials 4 and 7, effects were generally consistent across subgroups of
`gender, age, baseline disease severity, and race (although there were too few non-Caucasian
`subjects to draw meaningful conclusions regarding specific non-Caucasian subgroups). In both
`trials, the effect size tended to be more pronounced in subjects older than 40 years and subjects
`with more severe baseline disease.
`
`Efficacy Conclusions: I agree with Drs. Dinh, Lindberg, Levin, and Laughren that the dossier
`submitted by the applicant demonstrates efficacy for vilazodone, at a dose of 40 mg/d, for the
`treatment of MDD. The magnitude of the treatment effect is, at best, only modest, but the
`results are statistically persuasive, and appear robust to sensitivity analyses.
`
`
`
`
`
`
`
`
`
`
`Safety:
`
`Adequacy of Exposure: The principal support of safety came from the five phase 2 studies and
`two phase 3 studies. All were 8 weeks in length, and they included 1578 patients with MDD
`exposed to vilazodone. In addition, the applicant submitted a long-term open-label study,
`designed to meet the 1994 ICH E1 guidelines for drugs intended for long-term treatment of non-
`life-threatening conditions (which it barely did). The long-term study included 599 subjects
`overall (exposed to vilazodone), of whom 314 were exposed for 6 months (ICH E1: N should be
`300 to 600) and 118 were exposed for 1 year (ICH E1; N should be >100). Exposure in all
`studies, including the shorter phase 1 studies, totaled 2898 adult subjects exposed to one or
`more doses of vilazodone. The phase 2 and 3 studies included 309 subjects over 55 years old,
`and 37 subjects over 65 years old. There was no pediatric exposure in this program.
`
`Deaths: There were 3 deaths, but none of these subjects had received vilazodone at any time.
`
`Serious Adverse Events: Approximately ~100 serious adverse events (SAEs) were reported in
`81 subjects in the vilazodone development program. Many were psychiatric in nature, and
`probably represented worsening of the underlying condition – events typical and expected in
`psychiatric drug development programs.
`
`Reference ID: 2895106
`
`Page 10
`
`(b) (4)
`
`(b) (4)
`
`(b) (4)
`
`

`

`
`The applicant’s accounting of the SAEs did not reveal any patterns suggestive of vilazodone-
`related toxicity. Overall, the proportions of subjects with SAEs were similar in the vilazodone
`(1.8%) and placebo groups (2.3%). Dr. Lindberg, the primary clinical reviewer, carefully
`considered the narrative of each SAE, and found none of particular concern (except serotonin
`syndrome, see below). Most were relatively common types of background events, typically
`reported in clinical trials, with no patterns suggesting that any particular events, or cluster of
`related events, were more common in subjects exposed to vilazodone. Specifically, there were
`few adverse events that occurred in more than one subject in the vilazodone (or placebo) group.
`
` inspected the SAEs in this submission in the adverse event dataset (SAS transport file
`\0000\m5\datasets\iss \analysis\d-ae.xpt), blinded to treatment group, and coded each event as
`shown (Table 4):
`
`Table 4: My Coding and Summary of SAEs in the Phase 2 and 3 Trials (number of events, %)
`
`
` I
`
`Reference ID: 2895106
`
`Page 11
`
`

`

`N
`infection
`suicidal ideation, attempt, suicide
`pregnancy, positive pregnancy test
`pneumonia
`chest pain (not angina or unknown)
`diarrhea, colitis, enteritis, proctitis, gastroenteritis
`cholelithiasis, cholecystitis, gall bladder disorder
`overdose
`transient ischemic attack
`depression
`fever
`abscess, boil, furuncle
`cancer (non-squamous cell)
`all neoplasia
`angina
`arteriosclerosis, vascular disease
`deep venous thrombosis
`thrombophlebitis, thrombosis, thrombus, clot, embolism
`hypertension
`dehydration, volume depletion
`arrhythmia
`supraventricular arrhythmia
`atrial fibrillation
`hypokalemia
`hyponatremia
`tremor, shakiness, trembling
`psychosis, hallucinations
`anxiety, nervousness, panic attacks
`headache
`migraine
`blood urea nitrogen or creatinine elevated, acute/chronic renal failure
`bleeding
`apnea, respiratory failure, cyanosis
`COPD, COPD exacerbation
`pulmonary embolism
`weight gain
`serotonin syndrome
`bronchitis, tracheitis, bronchiectasis
`fracture
`urinary stone
`hernia
`wheeze, bronchospasm, asthma
`
`vilazodone
`2177
`8 (0.4)
`7 (0.3)
`6 (0.3)
`4 (0.2)
`3 (0.1)
`3 (0.1)
`3 (0.1)
`3 (0.1)
`2 (0.1)
`2 (0.1)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`1 (0)
`0 (0)
`0 (0)
`0 (0)
`
`placebo
`997
`1 (0.1)
`7 (0.7)
`5 (0.5)
`1 (0.1)
`0 (0)
`0 (0)
`0 (0)
`1 (0.1)
`0 (0)
`3 (0.3)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`0 (0)
`1 (0.1)
`2 (0.2)
`1 (0.1)
`1 (0.1)
`2 (0.2)
`
`
`In this classification/accounting system, various SAEs are classified under more than one
`heading, and various headings include more than one SAE. For example, I classify a single
`SAE of “atrial fibrillation” not only as “atrial fibrillation,” but also as an “arrhythmia,” and a
`“supraventricular arrhythmia.” Conversely, the heading “infection” includes “pneumonia,”
`“cholecystitis,” “abscess,” “bronchitis,” etc. Some isolated events that are uncommon and
`
`
`
`Reference ID: 2895106
`
`Page 12
`
`

`

`seemingly unlikely to be drug-related were not included, i.e., duodenal stricture, endometriosis,
`bronchoscopy, etc.
`
`Using this classification scheme, I found no patterns of concern here. None of the SAEs stand
`out as being more common in the vilazodone group, with the exception of pneumonia. Given
`this “signal,” one can consider the frequencies of infections in the common adverse events, and
`there was no difference between treatment groups (~19% for both vilazodone and placebo in
`the all-placebo-controlled safety database). This, in light of the lack of a reasonably plausible
`mechanistic explanation for infections, the small numbers of these SAEs, the multiplicity of
`events considered, and the lack of a “signal” for infections in the common adverse events, this
`single “signal” for pneumonia most likely represents play of chance and should be dismissed as
`such.
`
`Common Adverse Events:
`
`The common (non-serious) adverse events have been summarized by others. Excesses in
`gastrointestinal (GI) side effects (diarrhea, nausea, vomiting) predominate, and sleep disorders
`are also more common in the vilazodone group (Table 5):
`
`
`
`Gastrointestinal side effects: Gastrointestinal side effects were dose-limiting in phase 1 studies,
`and largely based on that experience, they are thought to be dose-related. It is remarkable that
`54% of subjects in the placebo-controlled phase 3 safety database experienced a
`gastrointestinal adverse event (source: applicant’s integrated summary of safety (ISS)
`[5.3.5.3.2], Table 31, page 107). Fortunately, there were relatively few discontinuations for GI
`side effects, and very few SAEs.
`
`The onset of GI side effects tended to occur in the first week or two of vilazodone treatment.
`This is well-illustrated in the applicant’s plots of time-to-first-event of nausea (Figure 3) and
`diarrhea (Figure 4):
`
`
`
`Reference ID: 2895106
`
`Page 13
`
`

`

`Figure 3: Time to Onset of Nausea (source: ISS appendix 2, Figure 1.3, page 17)
`
`
`
`Figure 4: Time to Onset of Diarrhea (source: ISS appendix 2, Figure 1.1, page 15)
`
`
`
`
`
`Reference ID: 2895106
`
`Page 14
`
`
`
`
`
`

`

`Sleep-related Adverse Events: Sleep disturbances were also of note, and were twice as
`common in subjects who received vilazodone as those who received placebo. In the all
`placebo-controlled safety database, sleep-related adverse events can be summarized as
`follows (Table 6):
`
`Table 6: Sleep-related Adverse Events in the All Placebo-controlled Safety Database, n (%)
`
`
`placebo
`997
`43 (4.3)
`22 (2.2)
`8 (0.8)
`
`
`
`vilazodone
`1578
`N
`insomnia 137 (8.7)
`abnormal dreams
`71 (4.5)
`nightmares
`30 (1.9)
`
`Reference ID: 2895106
`
`Page 15
`
`
`SSRI Reactions: Dr. Lindberg searched the safety database for cases suggestive of SSRI
`reactions, and found two subjects with probable serotonin toxicity. One occurred in a patient
`who had intentionally overdosed on vilazodone, and the other occurred in a patient during rapid
`titration of vilazodone.
`
`Ophthalmological Findings: There was concern regarding potential ophthalmological toxicity,
`based on an observation of transient corneal opacities (not lenticular opacities, i.e., cataracts) in
`a canine study. Longer-term animal studies did not confirm these findings, and found no
`evidence of cataract formation. Based on the initial animal findings, the phase 2 studies and
`longer-term phase 3 study included various types of ophthalmologic monitoring: slit lamp
`exams, fundoscopy, monitoring of intraocular pressure and visual acuity, and Schirmer’s test
`(lacrimation).
`
`The only finding from the phase 2 studies was a slight decrease in tear production, which may
`have accounted for some of the reports of blurred vision and the few reports of corneal
`abnormalities.
`
`In the controlled phase 3 studies, no treatment-emergent cataracts were reported in the listing
`of adverse events. Of 110 patients with a cataract at baseline in the 1-year open-label safety
`study, overall cataract severity was determined to have worsened for 14 patients (12.7%).
`There were also 4 patients in the 1-year study (3.6%) with apparently normal lenses at baseline,
`in whom new lens changes were reported. Other ophthalmologic adverse events in the open-
`label study included dry eye (4.7%), blurred vision (4.0%), and lacrimation increased (1.2%). All
`of these rates are difficult to interpret in the absence of a control group.
`
`Dr. Wiley Chambers, the ophthalmology consultant, noted in his final consultation (dated
`January 7, 2011) that the applicant has adequately assessed the ophthalmological safety profile
`of v