FDA Introductory Remarks Gastrointestinal Drugs Advisory Committee

NDA 021200, Zelnorm®: For the treatment of irritable bowel syndrome – constipation (IBS-C) in females <65 years

Preeti Venkataraman, MD Clinical Team Leader Division of Gastroenterology and Inborn Errors Products Office of Drug Evaluation III Center for Drug Evaluation and Research, FDA Product Overview

• Drug: tegaserod maleate

• Class: serotonin-4 (5-HT4) receptor agonist • Dose and mode of administration: 6 mg oral tablet twice a day • Proposed indication: treatment of women <65y with irritable bowel syndrome with constipation

2 Disease Overview

• Irritable bowel syndrome – constipation (IBS-C) is a functional GI disorder characterized by recurrent abdominal pain related to defecation, with hard or infrequent stools • Worldwide prevalence of IBS is ~11%, and the prevalence of IBS-C among the IBS subtypes is ~35% • Patients experience chronic symptoms with fluctuating severity and episodic flares • Pathophysiology is likely multifactorial, and underlying causes and clinical presentations can vary by patient • Main goals of treatment: improve abdominal pain, improve stool consistency, and increase frequency of bowel movements • There is a high prevalence of primary psychiatric disorders in IBS

3 FDA-Approved Treatments for IBS-C

FDA Year of IBS- Class Contraindications and Common Adverse Approved C Approval Events Treatments Chloride Contraindication: known or suspected mechanical GI Lubiprostone 2008 channel obstruction (Amitiza®) activator Common AEs: nausea, diarrhea, and abdominal pain Contraindications: patients with known or suspected Guanylate mechanical GI obstruction, patients <6 years of age due to Linaclotide 2012 cyclase-C risk of serious dehydration. (Linzess®) agonist Common AEs: diarrhea, abdominal pain, flatulence and abdominal distension. Contraindications: patients with known or suspected Guanylate Plecanatide mechanical gastrointestinal obstruction, patients less than 2018 cyclase-C 6 years of age due to the risk of serious dehydration. (Trulance®) agonist Common AEs: diarrhea 4 Regulatory History – Tegaserod (Zelnorm®)

• U.S. marketing approval – July 2002: For the short-term treatment of women with IBS-C – August 2004: For the treatment of men or women <65 years of age with chronic idiopathic constipation (CIC) • CV signal identification – February 2007: FDA notified of an imbalance in the number of CV ischemic events from a preliminary retrospective analysis of pooled clinical trial data – March 2007: Comprehensive analysis submitted and results of 1st external adjudication released – The rate of CV events in patients taking tegaserod was 13/11,614 (0.11%), versus 1/7,031 (0.01%) in patients taking placebo • U.S. marketing withdrawal – March 30, 2007: Novartis confirmed immediate suspension of marketing and sales, and a Public Health Advisory was issued • Suicidal Ideation/Behavior (SI/B) signal – February 2007: Labeling communication including the overall frequency of SI/B events – 8 patients out of 10,951 (0.07%) in the tegaserod group versus one patient out of 6,236 (0.02%) in the placebo group

5 Regulatory History – Tegaserod (Zelnorm®)

• 2 expanded access programs • April 2007: Emergency-IND program • October 2007: Treatment-IND program

• 2nd external adjudication – February 2008: Results from a reanalysis of CV ischemic events in 29 placebo-controlled trials were submitted

• FDA Advisory Committee meeting – November 2011: Discussion centered around recommendations for premarket CV safety development programs in the 5-HT4 receptor agonist class – Publicly available data regarding the risk of tegaserod were included in the discussion

Regulatory History – Tegaserod (Zelnorm®)

• FDA agreed that the proposed primary, secondary, and exploratory efficacy endpoints support consideration of reintroduction to the market. • Applicant was requested to define a population of severely symptomatic IBS-C patients – No final agreement on the severely symptomatic definition prior to submission of this efficacy supplement • FDA recommended that Novartis focus their reintroduction efforts on the IBS-C indication, and agreed with definition of a “low CV risk” population

7 Key Contents of NDA Submission

• Efficacy: Legacy data from 4 trials, 3 of which supported approval in 2002 [Studies 301, 358, and 307) – Trial 351 was not included in labeling as it was considered exploratory – FDA agreed to include this trial in reintroduction discussions in 2007, and it was included in the current post hoc evaluation of efficacy in a severely symptomatic IBS-C population • Safety: – Pooled databases (Db15, Db14) • Db15: 29 randomized, placebo-controlled clinical trials of ≥4 weeks’ duration and across multiple indications • Db14: 7 long-term (≥6 months), open-label studies – Adjudication reports and patient narratives – A non-interventional epidemiologic cohort study aiming to quantify the association between tegaserod use and cardiovascular outcomes – Nonclinical data, including data from QT and platelet aggregation studies

8 Benefit-Risk Considerations

9 Questions to Committee (1)

Question # 1 (Discussion): Discuss the strength of the potential CV safety signal of tegaserod, considering the totality of available data from clinical trials, adjudications, pharmacoepidemiology studies, nonclinical data, and pharmacovigilance data.

10 Questions to Committee (2)

Question #2 (Discussion): Discuss other potential safety concerns, including psychiatric safety adverse events of completed suicide and suicidal ideation/behavior, when considering reintroduction of tegaserod to the U.S. market.

11 Questions to Committee (3)

Question #3 (Voting): Is the reintroduction of tegaserod to the U.S. market supported by the available safety data? Discuss your answer.

12 Questions to Committee (4)

Question #4 (Voting): Do you agree that the therapeutic gain (treatment difference between tegaserod and placebo patients) is generally similar in magnitude between the severely symptomatic and originally approved population? Discuss your answer.

13 Questions to Committee (5)

Question #5 (Voting): In which patient population would you expect the benefits to outweigh the risks for patients treated with tegaserod? Choose from the following populations:

a. IBS-C females b. IBS-C females at low CV risk c. IBS-C females who are severely symptomatic d. IBS-C females at low CV risk and who are severely symptomatic e. Other

Discuss your answer.

Clinical Efficacy in Severely Symptomatic IBS-C Female Patients

Gastrointestinal Drugs Advisory Committee

Tegaserod Maleate October 17, 2018 Irena Lavine, MD, DGIEP Medical Officer Ling Lan, PhD, Statistician Outline

• Efficacy review strategy • Overview: – Clinical development program • Original approval: – Primary efficacy endpoint • Reintroduction: – Defining the severely symptomatic IBS-C subpopulation – Patient demographic and baseline characteristics of the severe subpopulation – Efficacy results in the severely symptomatic IBS-C subpopulation

17 Efficacy Review Strategy

• Original data which supported approval of tegaserod established efficacy in the female IBS-C subpopulation

• For reintroduction, post hoc analyses of completed trials in female patients with IBS-C were used to determine if a clinical benefit was observed in a severely symptomatic subpopulation

18 Overview: Clinical Development Program

• Trials supporting original approval – 301, 307, and 358 – Randomized, double-blind, placebo-controlled, multicenter trials – 4-week baseline period followed by a 12-week treatment period. Trial 358 had an additional 1-month withdrawal period – Different doses evaluated, but all trials included a 6 mg twice daily (BID) dose (FDA approved dose in 2002)

• Trial 351 was considered exploratory at the time of the original approval, but was included to support reintroduction because the same endpoints are now being evaluated in a post hoc nature – Similarly designed to Trial 301 with the same treatment arms and 12-week treatment duration

19 Original Approval: Primary Efficacy Results

Primary endpoint: responder defined as the overall IBS relief on subject global assessment (SGA): At least 50% of the SGAs with complete or considerable relief OR all (100%) with at least somewhat relief (i.e., complete, considerable or somewhat) over the last 4 weeks of treatment. Proportion of overall IBS relief responders during month 3, females only Trial Tegaserod 12 mg Placebo Percent Difference n/N (%) n/N (%) in Response (95% CI) 301 95 / 244 (39) 66 / 240 (28) 11.4 (3, 20) 307 100 / 233 (43) 88 / 234 (38) 5.3 (-4, 14) 358 334 / 767 (44) 292 / 752 (39) 4.7 (0, 10) 351 111 / 234 (47) 78 / 235 (33) 14.2 (6, 23)

Source: Adapted from Applicant’s Table 5.110.2 of Information Request response dated 6/15/2018 Abbreviations: CI, confidence interval 20 Reintroduction: Applicant’s Definition of the Severely Symptomatic IBS-C Subpopulation

Female patients with IBS-C reporting an average of: • 3 or more days per week with severe or very severe abdominal pain/discomfort; AND • 5 or more days per week with hard, very hard, or no stools.

21 Interpretation of the Severely Symptomatic Definition • Applying different rounding methods to determine which patients met the criteria resulted in subpopulations of varying severity • The review team considered 3 rounding methods to define the subpopulations: – Ceiling rounding (Applicant’s method) – 0.5 rounding – No rounding

22 Interpretation of the Severely Symptomatic Definition • As a reminder, the proposed definition of the severely symptomatic subpopulation includes 3 or more days per week with severe or very severe abdominal pain/discomfort • For example, a patient would qualify with an average of – 2.1 days per week of severe or very severe abdominal pain/discomfort because the ceiling method rounds 2.1 to 3 days per week (Applicant’s method) – 2.5 days per week of severe or very severe abdominal pain/discomfort because the 0.5 method rounds 2.5 to 3 days per week – 3 days per week of severe or very severe abdominal pain/discomfort because the no rounding method requires the patient to meet the criteria exactly as defined

23 Methodology for Assessing Number of Days With Severe Symptoms

Sample Sizes for Severely Symptomatic Female Subpopulations Based on Different Rounding Methods

Rounding Method Trial 301 Trial 307 Trial 358 Trial 351

Ceiling (Applicant’s method) 135 121 638 140

0.5 rounding 115 101 477 111

No rounding 74 70 346 87

Original approval sample size 484 467 1519 469

Source: Reviewer’s (Dr. Ling Lan’s) analysis

24 Patient Demographics of the Severely Symptomatic Subpopulation Trial 301 Trial 307 Trial 358 Trial 351 Demographics* N=74 N=70 N=346 N=87 Age, years Mean (SD) 41.5 (12.2) 43.3 (13.5) 40.2 (10.6) 43.7 (12.8) Median (min, max) 41.0 (17, 67) 43.0 (19, 83) 40.0 (18, 68) 43.0 (20, 81) Age category, n (%) <55 years 63 (85.1%) 59 (84.3%) 315 (91.0%) 70 (80.5%) 55 to <65 years 8 (10.8%) 6 (8.6%) 27 (7.8%) 10 (11.5%) BMI, kg/m² Mean (SD) 24.2 (4.4) 25.1 (4.9) 25.5 (5.2) 25.5 (5.1) Race, n (%) White 72 (97.3%) 60 (85.7%) 273 (78.9%) 77 (88.5%) Black 0 5 (7.1%) 53 (15.3%) 8 (9.2%) Other 2 (2.7%) 5 (7.1%) 20 (5.8%) 2 (2.3%) Country/region U.S. 3 (4.1%) 61 (87.1%) 346 (100%) 85 (97.7%) Non-U.S. 71 (95.9%) 9 (12.9%) 0 2 (2.3%) * Females only, no rounding Source: Adapted from Applicant’s submission, response to Information Request received 7/9/2018 Abbreviations: BMI, body mass index; SD, standard deviation 25 Baseline Characteristics of the Severely Symptomatic Subpopulation SGA of Severe or Very Severe Abdominal Pain/Discomfort Percentage of Patients With an Average of ≥4 Days Per Week With Severe or Very Severe Abdominal Pain/Discomfort at Baseline, n (%) (No Rounding) Trial Tegaserod Placebo Total 6 mg BID 301 22 (64.7%) 25 (62.5%) 47 (63.5%) 307 27 (64.3%) 19 (67.9%) 46 (65.7%) 358 132 (73.7%) 111 (66.5%) 243 (70.2%) 351 26 (70.3%) 36 (72.0%) 62 (71.3%)

Source: Applicant’s submission, response to Information Request received 7/9/2018 Abbreviations: BID, twice daily

26 Baseline Characteristics of the Severely Symptomatic Subpopulation SGA of Severe or Very Severe Abdominal Pain/Discomfort

Source: Reviewer’s plot, created using Applicant’s data submitted on 7/9/2018 in response to Information Request. 27 Baseline Characteristics of the Severely Symptomatic Subpopulation SGA of Bowel Habit Percentage of Patients With Average Number of Days Per Week With No Stools, No Rounding Trial Tegaserod 6 mg BID Placebo Total Number of days n (%) n (%) n (%)

301 ≥4 25 (73.5%) 34 (85.0%) 59 (79.7%) ≥5 13 (38.2%) 24 (60.0%) 37 (50.0%) 307 ≥4 29 (69.0%) 12 (42.9%) 41 (58.6%) ≥5 19 (45.2%) 7 (25.0%) 26 (37.1%) 358 ≥4 157 (87.7%) 148 (88.6%) 305 (88.2%) ≥5 116 (64.8%) 89 (53.3%) 205 (59.2%) 351 ≥4 29 (78.4%) 38 (76.0%) 67 (77.0%) ≥5 12 (32.4%) 22 (44.0%) 34 (39.1%) Source: Reviewer’s table, adapted using the Applicant’s response to Information Request received 7/9/2018 Abbreviations: BID, twice daily 28 Baseline Stool Frequency in the Severely Symptomatic Subpopulation

Source: Reviewer’s plot, based on Applicant’s Information Request response submitted on 7/9/2018 29 EFFICACY RESULTS

30 Efficacy Analyses in the Severely Symptomatic Subpopulation Ceiling Method

Primary Endpoint: Reanalysis in Female Severely Symptomatic Patients (Ceiling Method)a Trial Tegaserod 12 mg Placebo Difference in Response Original Approval n/N (%) n/N (%) 95% CI (%) Treatment Difference 95% CI (%) 301 25 / 67 (37.3) 16 / 68 (23.5) 13.8 (-1.6, 29.1) 11.4 (3.1, 19.8)

307 24 / 69 (34.8) 19 / 52 (36.5) -1.8 (-19.0, 15.5) 5.3 (-3.6, 14.2)

358 133 / 320 (41.6) 108 / 318 (34.0) 7.6 (0.1, 15.1) 4.7 (-0.2, 11.0)

351 33 / 64 (51.6) 29 / 76 (38.2) 13.4 (-3.0, 29.8) 14.2 (5.5, 23.0)

Source: Applicant’s Information Request response dated 6/15/2018, verified by Reviewer Abbreviations: CI, confidence interval a Analysis used the ceiling method (Applicant’s proposed method) to select severely symptomatic patient data. 31 Primary Efficacy Endpoint Comparisons: Three Rounding Methods and Original Approval Populations

Source: Reviewer’s plot, based on Applicant’s Information Request response submitted on 7/9/2018. 32 Efficacy Conclusions in the Severely Symptomatic Subpopulation

• Treatment effects for the primary endpoint were notably different for various severely symptomatic subpopulations determined by different rounding methods across the trials • Overall, treatment differences were in favor of tegaserod (compared with placebo) in all versions of the severely symptomatic subpopulation in Trials 301, 358, and 351, except for the no rounding population in Trial 358 • In Trial 307, placebo patients had higher response rates for all severely symptomatic subpopulations • Exploratory analyses based on a variation of the 2012 IBS Guidance also reported a treatment effect. Although there are limitations discussed in the FDA briefing book, the treatment effect numerically favored tegaserod using this endpoint in the female severely symptomatic subpopulation (ceiling method)

Nonclinical Safety Findings of Tegaserod

Gastrointestinal Drugs Advisory Committee

October 17, 2018 Ke Zhang, Ph.D. Division of Gastroenterology and Inborn Errors Products Tegaserod – Nonclinical Studies

• In vitro cardiac electrophysiology – Effect on the hERG channel – Effect on action potentials

• Binding affinity for multiple serotonin receptor subtypes

• Effects on the isolated coronary artery

• In vivo cardiovascular safety pharmacology

36 In vitro Cardiac Electrophysiology Studies

• hERG channel in HEK293 cells

– IC50 for Tegaserod = 13µM

– IC50 for Cisapride = 0.044µM • Guinea pig ventricular papillary muscle – No effect on action potentials up to 1µM • Human atrial myocytes – No effects on action potentials up to 0.1µM • Human plasma level (C ) ~0.01µM following 6 mg BID max 37

Binding Affinity for Multiple Serotonin Receptor Subtypes

• Tegaserod (HTF-919) is a 5-HT4 agonist with moderate affinity for 5-HT1 and 5-HT2 receptor subtypes

• M29, the major metabolite, and 2 minor metabolites have no binding affinity for 5-HT1B or 5-HT1D receptors • Human plasma level (C ) ~0.01µM or 10nM following 6 mg BID max 38

Isolated Coronary Artery Studies

• Tegaserod had no contractile activity in isolated coronary arteries from pigs (30μM), non-human primates (10μM), or humans (30μM)

• Tegaserod produced a small contractile response in canine coronary arteries at 3 to 10μM

Human plasma level (Cmax) ~0.01µM following 6 mg BID

39 Isolated Coronary Artery Studies (Human)

Source: Higgins, DL, MP Ero, M Loeb, K Kersey, A Hopkins, and DT Beattie, 2012, The inability of tegaserod to affect platelet aggregation and coronary artery tone at supratherapeutic concentrations, Naunyn Schmiedebergs Arch Pharmacol, 385(1):103-109. 40

Isolated Coronary Artery Studies (Animals)

In vivo Cardiovascular Safety Studies in Dogs

• Study with intraduodenal doses up to 10 mg/kg

• Study with oral doses up to 10 mg/kg

• No effects on blood pressure, heart rate, cardiac output, ECG (QT interval)

For reference:

Dog Cmax: 401 ng/ml for males at 10 mg/kg 277 ng/ml for females at 10 mg/kg

Human Cmax: ~3 ng/ml or 0.01µM following 6 mg BID

In vivo Cardiovascular Safety Studies (repeated dose toxicity studies in dogs)

• 2-week IV toxicity study at doses up to 1 mg/kg/day • 26-week oral toxicity study at doses up to 60 mg/kg/day • 52-week oral toxicity study at doses up to 70 mg/kg/day

No effect on ECG (heart rate and QT interval) No histopathological changes in heart

Clinical dose: 6 mg BID or 0.2 mg/kg/day (60 kg body weight assumed)

Tegaserod: Summary of Nonclinical Studies

• Tegaserod is a weak inhibitor of hERG potassium currents, but did not induce QT prolongation in in vivo studies in dogs • Tegaserod had no effects on action potentials in guinea pig papillary muscle or isolated human atrial myocytes

• 5-HT4 receptor agonist with moderate affinity for 5-HT1 and 5-HT2 receptor subtypes • Tegaserod, at clinically relevant concentrations, did not induce contraction in isolated coronary arteries from pigs, dogs, non-human primates, or humans

• In conclusion, from a nonclinical standpoint, tegaserod lacks clinically relevant cardiovascular effects

Clinical Pharmacology Findings of Tegaserod

Gastrointestinal Drugs Advisory Committee

October 17, 2018 Jie (Jenny) Cheng, PhD Division of Clinical Pharmacology III Office of Clinical Pharmacology Office of Translational Sciences Center for Drug Evaluation and Research Outline

• Pharmacokinetics of tegaserod and major metabolite – Absorption, distribution, metabolism, and excretion

• Intrinsic and extrinsic factors – Organ impairment – Drug-drug interaction

• Tegaserod and its major metabolite M29 effect on platelet aggregation

47 Pharmacokinetics

Absorption • Tmax 1 hour • Oral bioavailability is about 10% • Dose proportional increase in AUC (2 mg to 12 mg) • Food reduces AUC (40 to 60%) and Cmax (20 to 40%) Distribution • Protein binding ~98% • Vdss: 368±223 L after IV administration Metabolism • Metabolized mainly via two pathways o Presystemic acid hydrolysis o Direct glucuronidation • Major metabolite, M29 o 10-fold, and 16-fold higher AUC and Cmax for M29 than tegaserod Excretion • Tegaserod was undetectable in urine • One-third of the orally administered dose excreted in urine, primarily as metabolites • t1/2 ~4.6 – 8.1 hours Source: Approved Zelnorm USPI (https://www.accessdata.fda.gov/drugsatfda_docs/label/2007/021200s014lbl.pdf) and Study A2422 48 Effects of Intrinsic and Extrinsic Factors on PK

Hepatic impairment

Mild 31% higher AUC, 16% higher Cmax No dosage adjustment Moderate to severe Not studied Not recommended Renal impairment Mild to moderate Not studied No dosage adjustment

Severe • No difference in Cmax and AUC Not recommended • M29: 2-, and 10-fold higher Cmax and AUC, respectively Age

18-40 years vs. 65-85 years of age 22 % higher Cmax, 40 % higher AUC in No dosage adjustment (male and female) elderly female subjects Effects of other drugs on PK

P-gp inhibitor: • 63% higher Cmax Quinidine 600 mg (multiple dose) • 70% higher AUC0-12 49 Source: Approved Zelnorm USPI (https://www.accessdata.fda.gov/drugsatfda_docs/label/2007/021200s014lbl.pdf) and Study A2422 Effect of Tegaserod and M29 on Platelet Aggregation • Light Transmission Aggregometry for platelet aggregation – Light transmission through platelet rich plasma increases upon induction of platelet aggregation – Determine % platelet aggregation with or without tegaserod (or M29)

• For tegaserod: Blood from healthy subjects or IBS-C patients (2008)

– 10, 33, and 100nM for tegaserod; Cmax ~10nM after 6 mg BID

• For M29: Blood from healthy subjects (2017)

– 10, and 100nM for M29; Cmax ~160nM after 6 mg BID

50 Positive Effects of Tegaserod on % Platelet Aggregation Blood from 20 healthy subjects (5 males, 15 females) No positive control

Agonist Vehicle 10nM 33nM 100nM (10x Cmax) ADP 5µM 64.9±5.4 74.8±13.4** 75.4±11.2*** 79.3±11.0*** ADP 20µM 74.7±5.3 77.8±6.7 81.6±7.8** 87.1±7.6*** Collagen 1 µg/mL 75.1±5.2 77.1±6.4 82.5±6.9*** 90.3±7.3*** Collagen 5 µg/mL 81.3±7.0 84.9±7.7 88.7±8.3** 90.1±8.6** Epinephrine 5µM 82.7±6.9 85.6±8.5 90.0±6.6** 94.7±7.2*** Serotonin 5µM 15.3±4.4 21.6±4.8*** 22.0±6.2*** 26.7±7.5*** TRAP 20µM 88.3±1.4 88.6±1.9 89.0±1.5 88.9±3.2 Mean ± SD, *,**,***p < 0.05, 0.01, 0.001 compared to vehicle, respectively. Source: Adapted from Applicant’s table (Table 3-3, Study Report RD-2008-00298) 51 Positive Effects of Tegaserod on % Platelet Aggregation Blood from 20 IBS-C patients (6 males, 14 females) No positive control

Agonist Vehicle 10nM 33nM 100nM (10 x Cmax) ADP 5µM 66.0±6.1 73.6±12.0* 78.9±10.8*** 77.2±12.5*** ADP 20µM 74.6 ±5 .5 77.8±7.0 81.4±10.2* 86.3±7.2*** Collagen 1 µg/mL 75.8±5.5 79.7±5.8* 85.4±10.2** 88.1±10.4*** Collagen 5 µg/mL 80.0±7.1 85.0±8.6 86.5±8.0** 90.2±6.9*** Epinephrine 5µM 80.4±7.1 83.5±9.8 91.6±5.7*** 94.3±6.0*** Serotonin 5µM 13.5±2.8 20.5±5.2*** 20.7±4.4*** 22.5±7.2*** TRAP 20µM 88.4±1.5 88.0±1.9 89.4±1.7 88.6±2.2 Mean ± SD, *,**,***p < 0.05, 0.01, 0.001 compared to vehicle, respectively Source: Adapted from Applicant’s table (Table 3-4, Study Report RD-2008-00298) 52 Inconsistent Reports of Effect of Tegaserod on Platelet Aggregation • Higgins et al. (2012) • Blood from Healthy Subjects (n=10) • Agonists: ADP 5 µM; ADP 1 µM + 5-HT 5µM • No significant effects on platelet aggregation at tegaserod 10 nM, 33 nM and 100 nM • No positive control • Conlon et al. (2018) • Blood from healthy subjects (n=20, 7 males, 13 females) • Agonists: ADP 1.5 µM; TRAP 1µM • No significant effects of 100nM tegaserod on platelet aggregation • A positive control, TPO 100 ng/ml showed significant effects

Higgins, DL, MP Ero, M Loeb, K Kersey, A Hopkins, and DT Beattie, 2012, The inability of tegaserod to affect platelet aggregation and coronary artery tone at supratherapeutic concentrations, Naunyn Schmiedebergs Arch Pharmacol, 385(1):103-109. Conlon, K, JH De Maeyer, C Bruce, JAJ Schuurkes, L Christie, J McRedmond, K Derakhchan, and PR Wade, 2018, Nonclinical Cardiovascular Studies of Prucalopride, a Highly Selective 5-Hydroxytryptamine 4 Receptor Agonists, JPET, 364: 156-169. 53 Effect of M29 on Platelet Aggregation

Study USWM25JUN2017 (Blood from 20 healthy subjects; 11 males, 9 females) Vehicle M29 concentration Positive control

Agonist 10nM 100nM (0.6x Cmax) TPO (500 U/mL) ADP (5µM) 81.5±8.8 85.4±11.2 86.5±10.4 87.4±8.8 * Epinephrine (10µM) 71.8±26.7 72.0±26.4 83.0±21.6 88.5±10.6 * 5HT (5µM) + ADP (1µM) 63.6±17.3 66.6±17.3 70.7±17.5 86.2±10.5 *** Collagen (5 µg/mL) 86.6±9.5 87.3±10.5 87.9±8.7 87.7±10.4 TRAP (5µM) 85.4±8.7 87.7±11.0 85.6±11.3 86.7±10.9 EDTA (75mM) 1.4±1.8 1.2±1.8 0.8±1.1 1.0±1.6 Saline 7.1±4.6 8.2±6.8 6.4±3.9 20.2±27.6 * Mean ± SD, *,***p < 0.05, 0.001 compared to vehicle; TPO, thrombopoietin.

Source: Reviewer’s table, created using the Applicant’s data submitted on 08/29/2018 in response to Information Request. 54 Effect on Platelet Aggregation • Tegaserod • Mild but statistically significant concentration-dependent increase in platelet aggregation – Consistent results with blood from IBS-C patients and healthy subjects • Inconsistent results for tegaserod on platelet aggregation in published studies • M29 • No significant effects at 100nM; significant effects with positive control TPO • Inconclusive because effects were studied at concentrations lower than the therapeutic concentration 55 Summary

• Tegaserod is mainly eliminated by metabolism to form the major inactive metabolite, M29

• Overall, the in vitro results of tegaserod and M29 effects on platelet aggregation are inconclusive

Clinical Safety Evaluation

Gastrointestinal Drugs Advisory Committee

Tegaserod Maleate October 17, 2018 Sandhya Apparaju, Ph.D. Van Tran, Ph.D. Joel Weissfeld, M.D. Tegaserod Clinical Safety – Outline

• CV Ischemic (CVI) signal – Overview of the signal identification and adjudication processes – CVI signal, including the subset of major adverse cardiovascular event (MACE) – Baseline CV risk of cases and safety population – CVI signal in the proposed ‘low CV risk’ population – Other CV safety aspects – Post-marketing safety (CVI) • Neuropsychiatric safety signal (SI/B) – Post-marketing safety (SI/B)

59 CVI Signal Identification - Internal Adjudication

Pooled analysis of 18,645 patients from 29 clinical trials (Database, Db15) [n=11,614, tegaserod and n=7,031, placebo]

Manual and automated search for CVI events (coronary, cerebrovascular and other ischemic) 24 Initial Cases Tegaserod: 20, Placebo: 4

Novartis Internal Adjudication Patients with a CVI event Excluded Patients Tegaserod: 18 (0.15%) (‘Probably not’ an event) Placebo: 2 (0.03%) Tegaserod: 2, Placebo: 2

Drug: MI: 4, Angina Pectoris: Drug: Angina: 1, MI: 1 7, CAD/stenosis: 4, stroke: 2, Placebo: Angina: 1, vasoconstriction: 1 intermittent claudication: 1 Placebo: Angina: 1, cerebrovascular disorder: 1

60 Source: Reviewer’s figure, based on Applicant’s data submitted on 02/26/2018 CVI Signal Identification – 1st External Adjudication

24 CVI cases from the Initial Search Tegaserod: 20, Placebo: 4

Updated patient narratives Re-adjudication of cases

Patients with CVI Events Excluded Patients Tegaserod: 13 (0.11%) Tegaserod: 7 Placebo: 1 (0.01%) Placebo: 3

Drug: MI: 3, CV Death: 1, Drug: MI: 1, Angina: 2, Unstable Angina: 6 CAD/stenosis: 3, Stroke: 3 vasoconstriction: 1 Placebo: TIA: 1 Placebo: Angina: 2, Intermittent claudication: 1 MACE (Subset of CVI) Tegaserod: 7 (0.06%)

Source: Reviewer’s figure, based on Applicant’s data submitted on 02/26/2018 61 CVI Signal Identification – 2nd External Adjudication

Meta-analysis of 18,645 patients from 29 clinical trials database (Db15) [n=11,614, tegaserod and n=7,031, placebo]

Four different automated searches Source data retrieval; Pre-specified criteria for CVI/arrhythmic/CHF events 304 Adjudicated Cases (Includes the 24 Initial Cases) Tegaserod: 198, Placebo: 106

Patients with CVI Events Patients with Arrhythmias Total Excluded Patients Tegaserod: 7 (0.06%) Tegaserod: 11 (0.09%) Tegaserod: 158 (1.36%) Placebo: 1 (0.01%) Placebo: 5 (0.07%) Placebo: 96 (1.36%)

Drug: Death: 1, MI: 1, Excluded Initial CVI cases: 16 Unstable Angina: 3 Tegaserod: 13 Stroke: 2 Probably no event: 8 Placebo: TIA: 1 (Angina: 4, CAD: 2, MI: 2) Insufficient data: 3 MACE (Subset of CVI) (CAD: 2, stroke: 1) Congestive HF: 2 Tegaserod: 4 (0.03%) Placebo: 3 62 Source: Reviewer’s figure, based on Applicant’s data submitted on 02/26/2018 Summary of CV Ischemic Signal

Tegaserod Placebo (N=11614) (N=7031) Initial Search- Novartis Total CV Events 20 (0.17%) 4 (0.06%) Internal Adjudication CVI cases 18 (0.15%) 2 (0.03%) (Novartis) Coronary 16 (0.13%) 1 (0.01%) Cerebrovascular 3 (0.03%) 1 (0.01%) Major cases 11 (0.09%) 1 (0.01%) 1st External Adjudication CVI cases 13 (0.11%) 1 (0.01%) (Mt. Sinai) Coronary 10 (0.09%) 0 Cerebrovascular 3 (0.03%) 1 (0.01%) MACE (a subset of CVI) 7 (0.06%) 0 2nd External Adjudication CVI cases 7 (0.06%) 1 (0.01%) (DCRI) Coronary 5 (0.04%) 0 Cerebrovascular 2 (0.02%) 1 (0.01%) MACE (a subset of CVI) 4 (0.03%) 0

Source: Reviewer’s table, based on Applicant’s data submitted on 02/26/2018 63 Proposed ‘Low CV Risk’ Population

• Female IBS-C patients <65 years with a ‘low CV risk’, defined by the following contraindications: – A history of CV ischemic disease such as MI, stroke, TIA, or angina – More than one CV risk factor: hypertension, tobacco use, diabetes, hypercholesterolemia, age ≥55 years, or obesity

64 Patient Characteristics – Initial 24 CVI Cases

All patients (Db15) All females Female IBS-C Drug Placebo Drug Placebo Drug Placebo

(N=20) (N=4) (N=12) (N=3) (N=7) (N=2) Age Group 18 to <55 8 1 7 1 4 0 55 to <65 5 1 4 1 2 1 ≥65 years 7 2 1 1 1 1 History CVI disease 12 1 6 0 4 0 CV Risk Patients with >1 17 2 10 1 6 1 Factors (RFs) CV RFs Type of CV RF Active smoking 7 0 5 0 3 0 Hypertension 11 1 6 1 4 1 Hyperlipidemia 11 2 6 1 4 1 Diabetes 1 0 1 0 0 0 Age ≥55 years 12 3 5 2 3 2 Obesity 5 1 3 0 1 0

Source: Reviewer’s table, based on Applicant’s data submitted on 02/26/2018

• A similar trend in baseline CV disease/risk factors for MACE 65 Patient Characteristics – Safety Populations All Patients (Db15) All Females from Db15 Female IBS-C Patients Treatment Tegaserod Placebo Tegaserod Placebo Tegaserod Placebo N=11614 N=7031 N=10167 N=6154 N=4490 N=2148 Gender (n %) Female 10167 (88%) 6154 (88%) 10167 (100%) 6154 (100%) 4490 (100%) 2148 (100%) Age group (n %) ≥65 years 690 (6%) 475 (7%) 484 (5%) 337 (5%) 164 (4%) 93 (4%) 18 to <65 years 10919 (94%) 6554 (93%) 9678 (95%) 5815 (95%) 4322 (96%) 2055 (96%) History of Ischemic CV Disease (n %) 287 (2%) 168 (2%) 201 (2%) 113 (2%) 63 (1%) 37 (2%) CV Risk Factors (n %) Active Smoking 1355 (12%) 603 (9%) 1192 (12%) 534 (9%) 922 (21%) 422 (20%) Hypertension 2045 (18%) 1248 (18%) 1641 (16%) 995 (16%) 663 (15%) 314 (15%) Hyperlipidemia 2070 (18%) 1164 (17%) 1745 (17%) 984 (16%) 868 (19%) 413 (19%) Diabetes Mellitus 439 (4%) 284 (4%) 352 (3%) 233 (4%) 110 (2%) 69 (3%) Age ≥55 years 2337 (20%) 1513 (22%) 1869 (18%) 1225 (20%) 728 (16%) 382 (18%) Obesity (BMI >30) 1836 (16%) 1166 (17%) 1610 (16%) 1033 (17%) 684 (15%) 349 (16%) Number of CV Risk Factors (n %) CV Risk Factor =0 5501 (47%) 3432 (49%) 4983 (49%) 3103 (50%) 1968 (44%) 935 (44%) CV Risk Factor =1 3300 (28%) 1917 (27%) 2891 (28%) 1660 (27%) 1461 (33%) 675 (31%) CV Risk Factor >1 2813 (24%) 1682 (24%) 2293 (23%) 1391 (23%) 1061 (24%) 538 (25%) Age <65 + No History of Ischemic CV Disease + CV 8639 (74%) 5217 (74%) 7756 (76%) 4664 (76%) 3398 (76%) 1588 (74%) RiskSource: factor Adapted ≤1 from Safety Statistical Reviewer’s table 66 CVI/MACE in the ‘Low CV Risk’ Subgroup of Db15

Low CV Risk Females, All Patients, Db15 Females, Db15 Db15 Drug Placebo Drug Placebo Drug Placebo Adjudication Event Type (N=11614) (N=7031) (N=10167) (N=6154) (N=7756) (N=4664) Internal (Novartis) CVI Events 18 (0.15%) 2 (0.03%) 10 (0.1%) 1 (0.02%) 2 (0.03%) 0 1st External CVI Events 13 (0.11%) 1 (0.01%) 8 (0.08%) 1 (0.02%) 1 (0.01%) 0 (Mt. Sinai) MACE 7 (0.06%) 0 5 (0.05%) 0 0 0 2nd External CVI Events 7 (0.06%) 1 (0.01%) 4 (0.04%) 1 (0.02%) 1 (0.01%) 0 (DCRI) MACE 4 (0.03%) 0 3 (0.03%) 0 0 0 Source: Reviewer’s table, based on Applicant’s data submitted on 02/26/2018

• Note: ‘Low CV risk’ defined as females, age <65 years, without CV ischemic disease and with ≤1 CV risk factor (hypertension, hyperlipidemia, obesity, active smoking, age ≥55, diabetes mellitus) • Across all indications of Db15, narrowing the population resulted in a reduction of CVI events to 1, and MACE to 0 • There were no CVI events in low CV risk IBS-C patients • Patient level data support the safety of tegaserod in the restricted population 67

Other CV Safety Aspects of Interest

• BP: No clinically relevant changes at the proposed 6 mg BID dose; – SBP increase of up to 2 mmHg at doses >12 mg/day – per FDA risk analysis, a 10-year CV risk is unaffected in the proposed ‘low CV risk’ IBS-C females, following small to moderate BP increases • ECG: No formal TQT study was conducted for tegaserod; in clinical trials there were no meaningful effects on ECG intervals, including QTcF • Arrhythmia: No safety signal for clinically relevant arrhythmias with tegaserod use • Long-term Safety: 10 cases (4 CVI and 6 arrhythmias) were confirmed in the long- term use database – frequency of events comparable to that seen in placebo-controlled trials – longer duration of tegaserod use did not suggest greater frequency of CVI

68 Pharmacovigilance Review – CVI Events

In FDA Adverse Event Reporting System (FAERS): • 67 coronary (39) and cerebrovascular (28) events in females • 61 patients did not have a low CV risk – CVI disease history (35), and 2 or more CV risk factors (55) • 6 events (2 coronary, and 4 cerebrovascular) in ‘low CV risk’ – 3 of 4 patients with cerebrovascular events on concomitant estrogen or hormone replacement therapy • Limitations of FAERS data (e.g., missing information on CV risk factors, under- reporting of events)

Note that CVI disease and the presence of two or more CV risk factors are not mutually exclusive

69 CV Safety Signal – Summary

• Small incidence of CVI/MACE in tegaserod-treated patients • An imbalance for tegaserod vs. placebo remains • Difficult to interpret the signal, given its size relative to the safety database, as well as procedural/data limitations – Retrospective nature of analysis, incomplete source data retrieval, lack of objective data, differences across adjudication methods, etc. • Narrowing the population reduces the number of CVI events • A benefit-risk evaluation is important – Residual uncertainty as it relates to CVI signal – Availability of other treatment options for IBS-C

70 Neuropsychiatric Events – Signal Identification

• A higher incidence of SI/B events in tegaserod-treated patients of placebo-controlled clinical trials • 8/10,951 (0.07%) on drug vs. 1/6,236 (0.02%) on placebo – 1 additional case of completed suicide occurred in an open-label study Treatment Tegaserod Placebo N 12,281 7,564 Person-years 1881.5 1154.4 SI/B adverse events (total) 8 1 Rate/1000 person-years 4.3 0.9 Suicidal adverse events by category (n) Completed suicide 1 0 Suicide attempt 2 1 Self-injurious behavior, intent unknown 4 0 Suicidal ideation 1 0 Total 8 1

Source: Adapted from Dr. Andrew Mosholder’s consult review dated 10/24/2006

71 Neuropsychiatric Events

• A high prevalence of primary psychiatric disorders in IBS – Risk factors for SI/B – Prior psychiatric illnesses in cases of completed suicide – Suicidal ideation in patients receiving antidepressant medications • A high baseline frequency among IBS patients however, may not explain the treatment imbalance in drug vs. placebo – FDA recommended inclusion of language for SI/B in the Warnings & Precautions • Psychiatric events in Db15 – Overall: tegaserod 3.1% vs. placebo 2.5% – Most common AEs: insomnia (1.1% vs. 1%), anxiety (0.7% vs. 0.5%), depression (0.6% vs. 0.7%), nervousness (0.2% vs. 0.1%) – AEs occurred at similar rates in tegaserod vs. placebo 72 Pharmacovigilance Review – SI/B

• 5 completed suicides and 6 cases of suicidal ideation in tegaserod patients • All had a history of psychiatric disorders (one with prior SI/B) • No clear evidence of a causal relationship between SI/B and tegaserod use • Observational cohort study – No differences in suicide and self-injury incidence among tegaserod users vs. non-users – Limitation: insensitive method of reporting suicide

Cardiovascular Outcomes Meta-Analysis of Clinical Trials

Gastrointestinal Drugs Advisory Committee

October 17, 2018 Van Tran, Ph.D. Office of Biostatistics

75 FDA’s Approach to CV Assessment

• Objective: For each adjudication, compare CV risk in tegaserod exposed (all doses) patients vs. placebo • Meta-analysis features: – Synthesize information from all trials for more precise estimate of CV risk – Preserve within-study randomization by stratifying on trial – Include information from trials with zero CV events

76 Trial Description

• Includes all 29 randomized placebo-controlled trials: – 24/29 double-blind, multi-center, parallel group trials – 28/29 trials of <12 weeks duration – Number of patients per trial ranges 12 to 2657 subjects – Multiple indications and dosages – CV information assessed retrospectively – Patient level data

77 Statistical Methods

• Analysis population: all patients exposed to randomized treatment with at least one post-baseline safety evaluation • Trial set: all 29 randomized trials • Outcomes: MACE and ischemic events

78 Meta-Analysis Method

• Effect measure: risk difference – Measures excess/reduction in number of CV events per 10,000 patients in tegaserod arm compared to placebo • Estimator: Mantel-Haenszel risk difference stratified by trial • Model: fixed effects • α-level not adjusted for multiple testing

79 Patient Characteristics

Tegaserod Placebo Characteristic N=11,614 N=7,031 n (%) n (%) Sex Male 1447 (12) 877 (12) Female 10,167 (88) 6,154 (88) Age (yrs) Mean (SD), min-max 43 (13), 13-89 43 (14), 17-87 Race White 8,972 (77) 5,065 (72) Black 745 (6) 484 (7) Asian 670 (6) 669 (9) Other 1,227 (11) 813 (12) History of Ischemic Disease 287 (2) 168 (2) Cardiovascular Risk Factors Active smoking1 1,355 (12) 603 (9) Hypertension 2,045 (18) 1,248 (18) Hyperlipidemia 2,070 (18) 1,164 (17) Diabetes mellitus 439 (4) 284 (4) Age ≥55 years 2,337(20) 1,513 (22) Obesity (BMI >30)1 1,836 (16) 1,166 (17) Abbreviations: BMI, body mass index; SD, standard deviation 1 Patients without the condition are not distinguished from patients missing data for the condition in calculating percentages Source: Reviewer’s analysis 80 Meta-Analysis Results

Adjudication Endpoint Tegaserod Placebo Risk Difference1 N=11,614 N=7,031 (95% CI) Internal adjudication by Novartis Major cases: Any (December 2006) ischemic event2 11 1 7.6 (1.6, 13.7) 1st external adjudication by Mt. Sinai Confirmed (March 2007) ischemic event 13 1 10.1 (3.2, 17.0)

MACE 7 0 5.4 (1.1, 9.7) 2nd external adjudication by Duke Confirmed (May-October 2007) ischemic event3 7 1 3.8 (-0.8, 8.4)

MACE 4 0 3.1 (-0.1, 6.3) Abbreviations: CI, confidence interval; CVA, cerebrovascular accident; MACE, major adverse cardiovascular event; MI, myocardial infarction 1 events per 10,000 patients 2 Comprised of confirmed, unconfirmed, or probably not 3 One patient classified as having unstable angina as the leading event even though patient also had myocardial infarction Major cases included MI, unstable angina, cardiac death, CVA, stroke MACE consists of non-fatal stroke, non-fatal myocardial infarction, and cardiovascular death Source: Reviewer’s analysis 81 2nd External Adjudication: MACE

82 Source: Reviewer’s analysis

Conclusions

• Few CV events from trials with small to moderate number of patients • All adjudications: increased number of CV events per 10,000 patients in the tegaserod arm compared with placebo • Meta-analysis limited by: – Short trial durations – Low CV-risk population – Retrospective assessment of CV information

An Assessment of a Cohort Study of Tegaserod and Cardiovascular Events

Gastrointestinal Drugs Advisory Committee

October 17, 2018 Joel Weissfeld, M.D. Methods Overview

• U.S. database of insurance claims • Propensity-score-matched cohorts • September 2002 – December 2006 timeframe • N=52,229 matched patient-pairs, 11.8% men and 23.6% ≥55 years of age

86 Study Outcomes Identified in Hospital Claims and Confirmed by Clinician Review of Patient Chart Abstracts

• Cardiovascular Ischemic Event (CVIE) – Acute Myocardial Infarction (AMI) – Acute Coronary Syndrome (ACS) – Coronary Revascularization (CR) • Stroke

87 Tegaserod-Exposed Patients (N=52,229) With ≥1 Event, Fixed 6-Month Follow-up

Identified by Confirmed by Codes in Review of Outcome Claims Patient Charts CVIE 170 107 107 – 66 = 41 (38%) AMI 60 31 31 + 35 = 66 ACS 58 35 CR 107 83 Stroke 80 16

Source: Reviewer’s table, adapted from Applicant’s submission, Tables 2a and 2b in Final Study Report

88 Chart-Confirmed CVIE and Stroke Incidence, in 52,229 Matched Pairs, by Tegaserod Exposure Cohort

Incidence per 1000 Adjusted Hazard Event, patient-years Ratio Outcome EXP N Rate 95% CI HR 95% CI Cardiovascular YES 107 4.83 3.96, 5.83 0.95 0.73, 1.23 Ischemic Event NO 115 5.18 4.28, 6.22 Stroke YES 16 0.72 0.41, 1.17 0.90 0.46, 1.77 NO 18 0.81 0.48, 1.28 Source: Reviewer’s table, adapted from Applicant’s submission, Tables 6a-b in Final Study Report.

89 Results from as-Treated Analyses

• Mean duration of use, 2.4 months per exposed patient • CVIE: adjusted RR 1.14, 95% CI 0.83-1.56 • Stroke: adjusted RR 1.09, 95% CI 0.49-2.42

90 Interpretation of Results

• Confounding • Coronary Revascularization, no distinction between interventions for acute as opposed to chronic indications • Small number of events for Stroke

91 Conclusion

• Study-defined endpoints occurred no more frequently during 6-month post-index follow-up in tegaserod exposed than unexposed patients • Study assessed as sound (for a non-randomized study) • Study not regarded as comparable to a well-performed randomized trial with prospectively ascertained and rigorously adjudicated cardiovascular outcomes

FDA Backup Slides Shown CVI/MACE in ‘Low CV Risk’ Subgroups of Db15

Adj. Event All Patients, Db15 Females, Db15 Females at low CV Females at low CV risk (2-factor)a risk (3-factor)b Drug Placebo Drug Placebo Drug Placebo Drug Placebo (11614) (7031) (10167) (6154) (9548) (5748) (7756) (4664) Internal CVI 18 2 10 1 6 0 2 0 (0.15 %) (0.03 %) (0.1 %) (0.02 %) (0.06 %) (0.03 %) 1st Ext. CVI 13 1 8 1 4 0 1 0 (0.11 %) (0.01 %) (0.08 %) (0.02 %) (0.04 %) (0.01 %) MACE 7 0 5 0 2 0 0 0 (0.06 %) (0.05 %) (0.02 %) 2nd Ext. CVI 7 1 4 1 (0.02 2 0 1 0 (0.06 %) (0.01 %) (0.04 %) %) (0.02 %) (0.01 %) MACE 4 0 3 0 1 0 0 0 (0.03 %) (0.03 %) (0.01 %) a age < 65 years + no CVI history b age < 65 years + no CVI history + ≤ 1 CV risk factors CV risk factors: age ≥ 55 years, hypertension, hyperlipidemia, obesity, active smoking, diabetes

Source: Reviewer’s table, based on Applicant’s data submitted on 02/26/2018