CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`APPLICATION NUMBER:
`
`207620Orig1s000
`
`SUMMARY REVIEW
`
`
`
`
`
`
`
`
`
`
`
`
`
`

`

`Cross Discipline Team Leader Review
`
`Cross-Discipline Team Leader Review
`
`
`Date
`June 12. 2015
`
`From
`Aliza Thompson
`Sub'ect
`Cross-Disci line Team Leader Review
`NDA/BLA #
`NDA 207620
`A licant
`Novartis Pharmaceuticals Co oration
`
`Date of Submission
`PDUFA Goal Date
`
`December 17, 2015
`Au t 15. 2015
`
`Entresto / sacubitn'l and valsartan
`Proprietary Name / Established
`
`(USAN) names
`Dosage forms / Strength
`
`Fihn—coated tablets / strengths:
`24 mg sacubitlil and 26 mg valsartan
`49 mg sacubitn'l and 51 mg valsartan
`97 mg sacubitn'l and 103 mg valsartan
`
`Proposed Indication(s)
`
`
`
`Recommended:
`
`Approvalfor the treatment ofheartfailure pending resolution of
`the outstandin-
`roduct ua/itv issues and a . eement on lube/in '
`
`This secondary review is based on the following reviews:
`
`Quality (5/15/15)
`
`Wendy Wilson-Lee (Application Technical Lead):
`Anamitro Baneijee (Drug Substance). Sherita McLamore-
`Hines (Drug Product), Bogdan Kurtyka (Process). Robert
`Mello (Microbiology). Zhong Li (Facility). Salaheldin
`Hamed (Bio hannaceutics)
`
`Pharmacolo_ Toxicolo_ (5/15/15)
`
`William Link
`
`Clinical Pharmacolo _' (5/15/15)
`Clinical (5/ 15/ 15)
`Statistical (5/20/15)
`
`Risk Evaluation and Mitigation
`Strate _' (5/22/15)
`Division of Pediatric and Maternal
`Health 5/26/15
`
`Sreedharan Sabarinath. Lunin. Zhuan
`Kimberl Smith (Efficac ), Tzu-Yun McDowell (Safe )
`
`Somya Dunn
`
`Miriam Dinatale
`
`
`
`
`Division of Medication Error
`
`Janine Stewart
`
`Prevention and Analysis (6/ 1/15 and
`6/ 1 1/15)
`
`Patient Labelin_ 6/4/15
`
`Karen Dowd . Zama Patel
`
`(6/8/15)
`
`(6/8/15)
`
`Page 1 of 29
`
`Reference ID: 3778936
`
`1
`
`

`

`Cross Discipline Team Leader Review
`
`1. Introduction
`
`LCZ696 is a fixed—dose combination of valsartan and sacubitril. The proposed indication is as
`follows:
`(hm)
`
`There is widespread agreement among members of the review team that the submitted data
`support the efficacy and safety of the product for its intended use. There are, however,
`outstanding CMC issues and agreement needs to be reached with the applicant on labeling. In
`addition, Dr. Lawrence, the statistical reviewer, has raised concern that the application does
`not adequately address the Agency’s combination policy.
`
`2. Background
`
`Heart failure affects over 5 million patients in the United States; approximately half of these
`patients have heart failure with a reduced ejection fraction (HFrEF). HFrEF is associated with
`significant morbidity and mortality. Although a number of agents have been approved to treat
`HFrEF (most based on their effect on cardiovascular mortality and/or heart failure
`hospitalizations), there is still significant unmet need for therapies that can improve outcomes
`in these patients.
`
`LCZ696 is a fixed—dose combination of valsartan, an ARB, and sacubitril, a neprilysin
`inhibitor. Valsartan, as monotherapy, is indicated for the treatment of heart failure (NYHA
`class II-IV). According to the label, valsartan significantly reduced hospitalizations for heart
`failure in this population. Sacubitril is an NME. Although at present there is no approved
`neprilysin inhibitor, as discussed later in the review, there is some experience with this class of
`
`agents.
`
`develop LCZ696 as a treatment for chronic heart failure followed in October 2009. There were
`a number of discussions with Novartis over the course of the product’s development. Topics
`included the active comparator in the applicant’s phase 3 trial (choice of active comparator and
`dose), what the applicant would need to do to address the combination policy, and the use of
`M“). These same topics have also
`generated discussion during the review of the applicant’s NDA and hence are discussed in
`greater detail below.
`
`M" the IND to
`
`0
`
`The combination policv and the proposed active comparator. In June 2009, Novartis
`submitted a request for Special Protocol Assessment of their multi—center, randomized,
`double-blind phase 3 trial to evaluate the efficacy and safety of LCZ696 compared to
`enalapril in treating patients with chronic heart failure; the Agency responded with a No-
`Agreement Letter. In that letter, the Agency indicated that the trial would need to assess
`whether one of the components of the combination was sufficient for the entirety of the
`benefit. As an alternative to the proposed comparison with enalapril, the Division
`suggested an add-on study to evaluate whether sacubitril added to the benefit of valsartan.
`
`Page 2 of 29
`
`Reference ID: 3778936
`
`2
`
`

`

`Cross Discipline Team Leader Review
`
`The Agency also voiced concern that the proposed dose of the active comparator (10 mg
`bid of enalapril) was inadequate since labeling recommended titration to a higher dose.
`
`At a follow-up Type A meeting in August 2009, the Division stated that “. . .the issue of
`whether or not both components of LCZ696 contribute to the overall effect may or may not
`matter” and indicated that it would not matter if Novartis showed an effect on
`
`nonreversible events, such as mortality, myocardial infarctions, or strokes. During the
`meeting, Novartis asserted that a NEP inhibitor alone study would be ethically impossible
`and that there was evidence that NEP inhibitors alone would not be effective. The Division
`
`asked the sponsor to submit for review any data or literature supporting their assertion that
`sacubitril alone could not be the sole contributor to the effect of the combination product.
`The Division also indicated that if sacubitril vs.
`lacebo was studied on to of
`
`
`
`months after the Type A meeting, Novartis opened their IND with their phase 3 trial, a
`
`
`randomized, double-blind trial comparing LCZ696 to enalapril.
`
` In January 2010, members of the review Division and the Stud End ints and
`
`
`La e Develo ment Team met with Novartis to discuss the use of
`
` that captured the important symptom concepts that define chronic heart failure for the
`
`target population. The Agency also noted that, based on the submitted qualitative study
`data, the most important symptom concepts appeared to be shortness of breath, tiredness,
`swelling, and pain.
`
`Accordin to the minutes, the A enc stated that the
`
`
`
`In April 2014, Novartis notified the Agency that the Data Monitoring Committee for their
`phase 3 trial had recommended early closure for compelling efficacy. The Agency
`subsequently granted fast track designation, rolling review and priority review.
`
`
`
`Page 3 of 29
`
`Reference ID: 3778936
`
`3
`
`

`

`Cross Discipline Team Leader Review
`
`3. CMC
`
`According to Dr. Wilson, at this time, the overall product quality recommendation is pending
`completion of facilities inspections and evaluations. Agreement also needs to be reached on
`the dissolution specification.
`
`Drug Substance: Sacubitlil
`substance but are
`
`m4) and valsartan are designated as the regulatory drug
`(m4) under the applicant’s quality system.
`
`Chemical names and structure:
`
`Sacubitn'l: 4- { [( l S,3R)- 1 -([ 1 , l '-Biphenyl]-4-ylmethyl)—4-ethoxy-3-methyl-4-
`oxobutyl]amino}-4—oxobutanoic acid.
`Valsartan: N-Pentanoyl-N- {[2'-( lHtetrazol-S-yl)[ l , l '-biphenyl]-4-yl]methyl} -L-valine
`
`4
`
`Me
`
`Me
`
`H\ co,
`N
`
`,
`
`.
`...
`MO HH
`0 7’
`T
`0
`
`Me
`
`l
`
`‘
`
`Me
`-
`
`i“
`O
`
`1
`
`'
`
`-3Na'
`
`-2‘12H20
`
`0
`/ .~
`{1
`
`,
`
`co,
`
`N N
`
`N
`
`N
`
`Sacubitril
`
`Valsartan
`
`Figure 1: Structural Formula of Sacubitril Valsartan Sodium hydrate
`
`Drug Product: The drug product is a fixed—dose combination available in the following
`strengths: 24 mg of sacubitril and 26 mg of valsartan; 49 m 7 of sacubitril and 51 mg of
`valsartan; and 97 mg of sacubitril and 103 mg of valsartan.
`Inactive ingredients include
`microcrystalline cellulose, low-substituted hydroxypropylcellulose, crospovidone, magnesium
`stearate, talc, and colloidal silicon dioxide.
`
`Expiration Date and Storage Conditions: A 24-month drug product expiration date has been
`granted when stored at a controlled room temperature and protected from moisture in the
`intended container closure. The drug product is packaged in bottles and unit dose blister
`packages as described on pages 6 and 7 of the Quality Review.
`
`Facilities review/inspection: As noted above, facilities inspections have not been completed.
`
`Post-marketing agreements: The applicant has submitted comparability protocols supporting
`post-approval changes to 1) the drug product manufacturing site, control, batch size and
`process and 2) the
`M“) intermediate manufacturing site, control, batch size and
`process. According to the Quality Review, these protocols are acceptable.
`
`2 During the product’s development and in the published trials, 24 mg/26 mg, 49 mg/51 mg and 97 mg/103 mg
`was referred to as LCZ696 50 mg, 100 mg and 200 mg. respectively.
`
`Page 4 of 29
`
`Reference ID: 3778936
`
`4
`
`

`

`Cross Discipline Team Leader Review
`
`4. Nonclinical Pharmacology/Toxicology
`According to Dr. Link’s review, the preclinical toxicology program was well-conducted and
`thorough and the application can be approved from a pharmacology/toxicology perspective.
`
`Toxicology:3 Target organs for LCZ696 and/or sacubitril toxicity included the kidney, red
`blood cells, heart and gastrointestinal tract. In brief, findings included: renal juxtaglomerular
`hypertrophy/hyperplasia in the rat and monkey and renal tubular changes (tubular basophilia,
`cytoplasmic vacuolation and single cell necrosis) in the monkey; reversible changes in
`hematology parameters (e.g., decreases in red blood cell count, hemoglobin concentration,
`hematocrit, and reticulocytes); decreased heart weights without histopathological findings; and
`reversible microscopic changes of focal glandular stomach mucosal erosion and mixed cell
`inflammation in rats and emesis and diarrhea without histologic correlates in the cynomolgus
`monkey. According to Dr. Link, these findings reflected adaptive responses or resulted from
`exaggerated pharmacodynamic responses to high doses and do not raise concern for safe use in
`humans. See pages 240 to 242 of Dr. Link’s review for further discussion of these findings.
`Genotoxicity and carcinogenicity: According to the Carcinogenicity study review, sacubitril
`had no effect on survival or tumor incidence in mice or rats. In genetic toxicity studies of
`LCZ696, sacubitril, and the active metabolite LBQ657, there was no evidence of genotoxic
`potential.
`Reproductive toxicology: LCZ696 had no effect on fertility in rats. In embryo-fetal
`development studies in rats and rabbits, treatment with LCZ696 during organogenesis resulted
`in increased embryo-fetal lethality at clinically relevant doses. Teratogenicity (i.e., a low but
`dose-dependent increase in the incidence of hydrocephaly) occurred in rabbits administered
`maternally toxic doses of LCZ696. Pre- and postnatal development studies with valsartan and
`sacubitril indicate that exposure to LCZ696 during these periods could impair fetal
`development and survival.
`Other notable issues: NEP is a major beta amyloid-degrading enzyme in the brain. Hence,
`there is a theoretical risk that NEP inhibition by LCZ696 could lead to accumulation of beta
`amyloid in the brain. To address this issue, the applicant assessed the effects of LCZ696 on
`amyloid-β concentrations in cerebrospinal fluid (CSF) and brain tissue in cynomolgus
`monkeys treated with a clinically relevant dose (50 mg/kg/day) of LCZ696 for 2 weeks.
`Treatment was associated with increases in Aβ 1-40, 1-42, and 1-38 levels in the CSF, without
`corresponding increases in Aβ levels in the brain. In a toxicology study in which cynomolgus
`monkeys were treated with 300 mg/kg/day of LCZ696 for 39 weeks (AUC exposure
`~2X the maximum recommended human dose), there was no amyloid-β accumulation in the
`brain.
`5. Clinical Pharmacology/Biopharmaceutics
`According to the Clinical Pharmacology Review, the application can be approved from a
`clinical pharmacology perspective.
`
`
`3 Per Dr. Link’s review, the pharmacologic targets of LCZ696 (the AT1 receptor and neprilysin) are
`evolutionarily conserved across mammalian species. Although there are some species differences in the rate of
`hydrolysis of sacubitril to LBQ657, all species are exposed to the same major compounds delivered by LCZ696.
`
`Page 5 of 29
`
`Reference ID: 3778936
`
`5
`
`

`

`Cross Discipline Team Leader Review
`
`Mechanism of action: LCZ696 inhibits neprilysin via LBQ657, the active metabolite of
`sacubitril, and blocks the angiotensin II type-1 receptor via valsartan. Effects of the renin-
`angiotensin-aldosterone system include vasoconstriction, renal sodium and fluid retention,
`activation of cellular growth and proliferation, and subsequent maladaptive cardiovascular
`remodeling. Neprilysin inhibition blocks neprilysin-dependent proteolytic degradation of
`natriuretic peptides, thus enhancing the effects of natriuretic peptides. Effects of natriuretic
`peptides include promoting vasodilation, natriuresis and diuresis, increased glomerular
`filtration rate and renal blood flow, inhibition of renin and aldosterone release, reduction of
`sympathetic activity, and anti-hypertrophic and anti-fibrotic effects.
`
`Pharmacokinetics:
`Absorption: Following oral administration, LCZ696 dissociates into sacubitril and valsartan.
`Absolute bioavailability of sacubitril from LCZ696 is estimated to be ≥ 60%. The absolute
`bioavailability of valsartan from LCZ696 is about 50% higher than the bioavailability of
`valsartan administered alone. Food does not have a clinically significant impact on systemic
`exposure.
`Metabolism: Sacubitril undergoes metabolism via esterases to form the active moiety,
`LBQ657. LBQ657 is not metabolized further into any major metabolites. Valsartan does not
`undergo significant metabolism.
`Distribution: Sacubitril, LBQ657, and valsartan are highly bound to plasma proteins (94-97%).
`Elimination: Approximately 52% to 68% of sacubitril is excreted in the urine (primarily as
`LBQ657), with the remainder excreted in the feces. Approximately 83% and 13% of valsartan
`is excreted in the feces and urine, respectively.
`Intrinsic factors affecting elimination: Age, sex, and body weight do not have a significant
`impact on exposure or efficacy outcomes. The effect of renal and hepatic impairment is
`discussed below.
`• Renal impairment: Steady state pharmacokinetic parameters of sacubitril and valsartan
`were similar in patients with mild, moderate, and severe renal impairment and matched
`healthy subjects. In subjects with mild (CrCl 50 to < 80 mL/min), moderate (CrCl 30 to <
`50 mL/min), and severe renal impairment (CrCl < 30 mL/min), steady state exposure of
`LBQ657 was 2.1 times, 2.2 times, and 2.7 times, respectively, that seen in healthy subjects.
`No study was conducted in patients undergoing dialysis. LBQ657 and valsartan are
`unlikely to be removed by dialysis as they are highly bound to plasma protein.
`• Hepatic impairment: The exposures (AUC) of sacubitril, LBQ657 and valsartan increased
`by ~53%, 48%, and 19%, respectively in mild hepatic impairment relative to healthy
`subjects. The exposures (AUC) of sacubitril, LBQ657 and valsartan increased by ~245%,
`90%, and 109% respectively in moderate hepatic impairment relative to healthy subjects.
`No study was conducted in patients with severe hepatic impairment.
`Drug-drug interactions: CYP450 enzyme-mediated metabolism of sacubitril and valsartan is
`minimal, hence drugs that impact CYP450 enzymes are not expected to impact LCZ696
`exposure. From a safety perspective, the main concern is for a pharmacodynamic as opposed
`to a pharmacokinetic interaction (i.e., angioedema with concomitant or proximate
`administration of an ACE inhibitor). Angiotensin converting enzyme degrades bradykinin and
`elevated bradykinin activity is thought to play a role in ACE inhibitor-induced angioedema.
`
`Page 6 of 29
`
`Reference ID: 3778936
`
`6
`
`

`

`Cross Discipline Team Leader Review
`
`Although not the major pathway for bradykinin degradation, neprilysin also degrades
`bradykinin. As discussed elsewhere in the review, the current label contains a contraindication
`against concomitant use with an ACE inhibitor or within 36 hours of switching from or to an
`ACE inhibitor.
`
`Pharmacodynamic effects:
`β-Amyloid concentrations: In healthy subjects, administration of LCZ696 400 mg once a day
`for 2 weeks was associated with a 42% increase in CSF β-Amyloid 1-38 relative to baseline
`and a 50% increase in plasma β-Amyloid 1-40. As noted in the clinical review, the clinical
`significance of these findings is not known.
`
`QT effects: No significant QTc prolongation was observed with LCZ696 (400 mg and 1200 mg)
`in a thorough QT study.
`
`Proposed dosing regimen and dosing rationale:
`The proposed dosing regimen (starting dose of 49 mg/51 mg twice-daily doubled after 2 to 4
`weeks to the target maintenance dose of 97 mg/103 mg) is based on the dosing regimen used
`in the LCZ696 run-in phase of the applicant’s phase 3 trial. The maximum dose in the phase 3
`trial was chosen based on an Emax model that showed no additional neprilysin inhibition (seen
`as an increase in cGMP) when the dose of LCZ696 was increased beyond 97 mg/103 mg per
`day; a twice-daily regimen was adopted to prolong the duration of cGMP elevation to 24
`hours. The proposed target dose also provides systemic exposure to valsartan similar to that
`from 160 mg valsartan BID, the approved target maintenance dose of valsartan for heart
`failure. As discussed in the Clinical Pharmacology Review, a dose/exposure response analysis
`of the phase 3 trial was not feasible because pharmacokinetic data were collected in only 7%
`of subjects and the trial employed dose titration based on tolerability.
`A lower starting dose (24 mg/26 mg twice-daily) is being recommended for patients not
`currently taking an ACE inhibitor or ARB and for patients on low doses of these agents. These
`populations were not included in the phase 3 trial as entry criteria specified a protocol-defined
`minimum dose of these agents, and entry into the LCZ696 run-in period required successful
`completion of the enalapril run-in phase. The recommendation to initiate these patients on a
`lower starting dose is based in part on the findings in a 12-week study in patients with HFrEF.
`The trial included a 5-day run in period in which all subjects were treated with 24 mg/26 mg of
`LCZ696 twice-daily. Following the run-in phase, subjects were randomized to 49 mg/51 mg
`twice-daily titrated to 97 mg/103 mg twice-daily or 24 mg/26 mg twice daily titrated to 49
`mg/51 mg twice-daily then 97 mg/103 mg twice-daily (see Figure 5 of the Clinical
`Pharmacology Review). In this study more subjects who were naïve to previous ACE
`inhibitor or ARB therapy or on low dose therapy were able to achieve and maintain the target
`dose when LCZ696 was up-titrated using the more gradual regimen.
`Based on pharmacokinetic and/or safety considerations, a lower starting dose (24 mg/26 mg
`twice-daily) is also being recommended in patients with severe renal impairment and moderate
`hepatic impairment. Since the dose can be titrated based on tolerability and because the effect
`of renal impairment differs by component, the target maintenance dose will remain 97 mg/103
`mg in these populations.
`
`
`Page 7 of 29
`
`Reference ID: 3778936
`
`7
`
`

`

`Cross Discipline Team Leader Review
`
`6. Clinical Microbiology
`The product is not an antimicrobial. According to the Quality Review, the tests and proposed
`acceptance criteria for microbial burden are adequate for assuring the microbial quality of the
`drug product.
`
`
`7. Clinical/Statistical- Efficacy
`Overview of PARADIGM-HF
`In support of the proposed indication, the applicant conducted a single phase 3 trial,
`PARADIGM-HF. PARADIGM-HF was an international, randomized, double-blind, double-
`dummy, parallel group, event-driven, active-controlled trial comparing LCZ696 with enalapril
`in 8,442 patients age ≥ 18 with NYHA class II to IV chronic heart failure and an LVEF ≤ 40%
`(≤ 35% per protocol amendment 1). The trial’s primary endpoint was the time to first heart
`failure hospitalization or CV death. Secondary endpoints included:
`time from randomization to all-cause death;
`•
`• change from baseline (randomization visit) in the clinical summary score for heart
`failure symptoms and physical limitations (as assessed by KCCQ) at 8 months;
`time from randomization to new onset of atrial fibrillation4;
`time from randomization to first occurrence of: a 50% decline in eGFR relative to
`baseline; a >30 mL/min/1.73 m2 decline in eGFR relative to baseline to a value below
`60 mL/min/1.73 m2; or ESRD.
`
`•
`•
`
`
`To be eligible for enrollment, patients had to have been on an ACE inhibitor or ARB at a
`stable dose of at least 10 mg/day of enalapril or a protocol-defined equivalent agent and on a
`stable dose of a beta-blocker (unless contraindicated or poorly tolerated) for at least four weeks
`before screening. After discontinuing their existing ACE inhibitor or ARB therapy, patients
`entered sequential single-blind run-in periods during which they received enalapril (10 mg
`twice-daily), followed by LCZ696 100 mg (equivalent to 49 mg sacubitril/51 mg valsartan)
`twice-daily, increasing to 200 mg (equivalent to 97 mg sacubitril/103 mg valsartan) twice-
`daily. Subjects who successfully completed the sequential run-in periods were randomized to
`receive either LCZ696 or enalapril 10 mg twice-daily (see figure below).
`
`
`
`4 Time from randomization to new onset of atrial fibrillation was added as a secondary endpoint in Protocol
`Amendment 3. Although somewhat of a late addition to the protocol (at the time of the amendment, ~64% of
`primary endpoint events had accrued), as discussed later in the review, no effect was seen on this endpoint.
`
`Page 8 of 29
`
`Reference ID: 3778936
`
`8
`
`

`

`Cross Discipline Team Leader Review
`
`
`
`Figure 2: Paradigm HF Study Design
`Source: Figure 9-1, Clinical Study Report for PARADIGM HF
`
`
`PARADIGM-HF was designed to have 80% power to detect a 15% reduction in
`cardiovascular mortality based on the following assumptions: a 7% annual cardiovascular
`death rate in the enalapril arm, an enrollment period of 18 to 22 months, and a minimum
`follow-up duration of 21 months. Based on the projected sample size (7980 patients to obtain
`1,229 cardiovascular deaths) and assuming a 14.5% event rate for the primary composite
`endpoint in the enalapril arm, 2410 composite endpoint events were expected, giving the trial
`97% power to detect a 15% reduction in the primary composite endpoint.
`
`The statistical analysis plan specified three interim analyses to assess for efficacy at 1/3, 1/2,
`and 2/3 of primary endpoint events (approximately 804, 1205 and 1607 patients, respectively,
`with a primary endpoint event). The Haybittle-Peto type of boundary was used to assess
`superiority with an alpha of 0.0001 (one-sided) spent at the first interim analysis and 0.001
`(one-sided) at the second and third interim analyses. Both the primary composite endpoint and
`the cardiovascular death component would need to meet the boundary to terminate the trial
`early for efficacy.
`
`According to the statistical analysis plan, the primary composite endpoint would be analyzed
`using a Cox regression model with terms for treatment and region and that the primary
`efficacy analysis was to include all positively adjudicated events occurring between
`randomization and March 31, 2014 (the date the trial was terminated early for efficacy). The
`analysis would be based on the full analysis set, which was defined as all randomized patients
`excluding misrandomized patients who did not qualify for randomization but were
`inadvertently randomized and did not receive study drug. 5
`
`If the primary endpoint was statistically significant, then the secondary endpoints would be
`tested using the same alpha as used for the primary endpoint at whatever time point the study
`was stopped. Testing of the four secondary endpoints would be done using a sequentially
`
`5 The statistical analysis plan also stated that “Further exclusions could be justified in exceptional circumstances
`(e.g., serious GCP violations),” and ultimately, additional subjects were excluded for such violations.
`
`Page 9 of 29
`
`Reference ID: 3778936
`
`9
`
`

`

`Cross Discipline Team Leader Review
`
`rejective multiple test procedure. As shown in Figure 1 of Dr. Lawrence’s review, the alpha
`was initially split between the first two secondary endpoints, with 0.8α allocated to all-cause
`mortality and 0.2α allocated to the KCCQ. If both hypotheses were rejected, the full alpha was
`to be allocated to the next secondary endpoint on the testing chain; if only one of the initial
`hypotheses was rejected, the alpha allocated to the rejected hypothesis would then be allocated
`to the next secondary endpoint.
`
`Results
`Demographics
`The baseline demographics of subjects who were randomized into the double-blind treatment
`period are shown in Tables 9-12 of the Clinical Review. As a whole, the two treatment arms
`were well-matched with regard to baseline characteristics. The mean age was 64 years and the
`majority of subjects were male (78%). Approximately 5% of subjects were enrolled at sites in
`the United States and approximately 5% of randomized subjects were black.
`
`Most subjects (70%) were NYHA Class II, 24% were NYHA Class III and fewer than 1%
`were NYHA Class IV. Mean ejection fraction was ~29%, mean baseline eGFR was 68
`mL/min/1.73m2, mean systolic blood pressure was in the low 120’s and mean heart rate was in
`the low 70’s. The majority of subjects were taking beta-blockers (94%), mineralocorticoid
`antagonists (58%), and diuretics (82%) at baseline. Overall, only 15% of subjects had an
`implantable cardioverter-defibrillator (ICD) or cardiac resynchronization therapy-defibrillator
`(CRT-D); however device use was reported in 60% of subjects enrolled at sites in the United
`States.
`
`Disposition
`A total of 10,521 subjects entered the run-in period. Of these, 1,102 subjects failed the
`enalapril run-in period and 982 failed the LCZ696 run-in period (10.5% and 10.4% of subjects
`entering the respective run-in periods). As shown in Table 15 of the Clinical Review, the most
`common reason for failing the enalapril and LCZ696 run-in periods was an adverse event,
`most often related to renal dysfunction, hyperkalemia, or hypotension.
`
`The disposition of subjects in the double-blind treatment period is shown in the table below.
`Overall, follow-up was adequate to assess the key efficacy findings. Approximately 17% of
`subjects prematurely discontinued therapy in the LCZ696 arm as compared with 19% in the
`enalapril arm. Vital status was unknown in 20 (0.2%) subjects. The most common reason for
`treatment discontinuation during the double-blind treatment period was an adverse event
`(12.1% of subjects randomized to enalapril and 10.4% of subjects randomized to LCZ696).
`
`
`Page 10 of 29
`
`Reference ID: 3778936
`
`10
`
`

`

`Cross Discipline Team Leader Review
`
`Table 1: Subject disposition during the randomized double-blind period
`
`Enalapril
`LCZ696
`n (%)
`n (%)
`4233 (100)
`4209 (100)
`Randomized
`4 (0.1)
`6 (0.1)
` Not treated
`4212 (99.5)
`4187 (99.5)
`Primary efficacy population (full analysis set)
`21 (0.5)
`22 (0.5)
` Excluded
` Misrandomized1
`2 (0.1)
`4 (0.1)
`19 (0.5)
`18 (0.4)
` Site excluded for GCP violations
`3379 (79.8)
`3441 (81.8)
`Completed study on treatment
`2869 (67.8)
`3011 (71.5)
` Alive at study termination
`815 (19.3)
`729 (17.3)
`Prematurely discontinued study treatment
`18 (0.4)
`17 (0.4)
`Did not complete study
` Withdrew consent
`13 (0.3)
`15 (0.4)
`4 (0.1)
`9 (0.2)
` Vital status unknown
`5 (0.1)
`2 (0.1)
` Lost to follow-up (vital status unknown)
`1Subjects who failed the run-in period for whom IVRS randomization calls were erroneously performed but who never
`received study medication.
`Source: Table 16, Clinical Review.
`
`Primary endpoint findings
`The trial was stopped for efficacy at the third interim analysis. As of the analysis cut-off date,
`the primary endpoint (death from cardiovascular causes or hospitalization for heart failure) had
`occurred in 914 subjects (21.8%) in the LCZ696 arm and 1117 patients (26.5%) in the
`enalapril arm (HR of 0.80; 95% CI 0.73; 0.87; 1-sided p=0.0000002). The Kaplan-Meier plot
`of the time to first event showed continued separation of the curves over time. As shown in the
`table below, the treatment effect reflected a reduction in both components of the composite.
`The results for the primary composite endpoint were for the most part consistent across the
`subgroups examined.
`
`
`Page 11 of 29
`
`Reference ID: 3778936
`
`11
`
`

`

`Cross Discipline Team Leader Review
`
`
`
`Figure 3: Kaplan-Meier plot for the primary composite endpoint (CV death or HF
`hospitalization)
`Source: Figure14.2-1.2, Clinical Study Report for PARADIGM HF
`
`Table 2: Primary Composite Endpoint (CV death or HF Hospitalization)
`Enalapril
`LCZ696
`
`Hazard Ratio
`(N=4212)
`(N=4187)
`(95% CI; 1-sided p-value)
`n (%)
`n (%)
`
`
`
`1117 (26.5)
`914 (21.8)
`0.80 (0.73, 0.87; 0.0000002)
`459 (10.9)
`377 (9.0)
`658 (15.6)
`537 (12.8)
`
`
`693 (16.5)
`558 (13.3)
`658 (15.6)
`537 (12.8)
`
`
`0.80 (0.71, 0.89)
`0.79 (0.71, 0.89)
`
`Primary Composite
`Endpoint*
` CV Death
` HF Hospitalization
`Subjects with events at
`any time**
`CV Death
`HF Hospitalization
`*Analysis shows time to first component; ** Analyses of the components of the primary composite endpoint were
`not prospectively planned to be adjusted for multiplicity.
`Source: Tables 18 and 19, Clinical Review
`
`Secondary efficacy endpoints
`Mortality:
`Since the trial met its primary endpoint, all-cause mortality was tested at the allotted alpha
`level (0.8 of the alpha used to test the primary endpoint). As shown in the table below, there
`were fewer deaths in the LCZ696 arm (HR of 0.84; 95% CI 0.76, 0.93; 2-sided p=0.0009).
`This finding was driven entirely by a lower incidence of cardiovascular mortality in subjects
`randomized to LCZ696.
`
`
`Page 12 of 29
`
`Reference ID: 3778936
`
`12
`
`

`

`Cross Discipline Team Leader Review
`
`Table 3: All-cause death
`
`
`LCZ696
`Enalapril
`(N=4187)
`(N=4212)
`n (%)
`n (%)
`711 (17.0)
`835 (19.8)
`All-cause death
`558 (13.3)
`693 (16.5)
`Cardiovascular death
`120 (2.9)
`109 (2.6)
`Non-cardiovascular death
`33 (0.8)
`33 (0.8)
`Unknown
`Source: Table 24, Clinical Review and Figure 5, Statistical Review
`
`Hazard Ratio
`(95% CI; 2-sided p-value)
`
`0.84 (0.76, 0.93; 0.0009)
`
`
`
`
`
`KCCQ Clinical Summary Score:
`The change in the KCCQ Clinical Summary Score from baseline (randomization) to month 8
`was also tested at its allocated alpha level (0.2 of the alpha used to test the primary endpoint).6
`At month 8, there was less of a decline in the Clinical Summary Score in the LCZ696
`treatment arm, as compared to the enalapril arm, however the treatment effect was small- the
`least square mean of the difference was 1.6 (95% CI 0.6, 2.7) in subjects with a mean baseline
`score of ~76 (on a scale of 0 to 100). Although the p-value for this analysis was quite low, the
`endpoint did not meet its pre-specified alpha threshold.
`
`Table 4: KCCQ Clinical Summary Score
`LSM of difference1
`
`LCZ696
`Enalapril
`(95% CI; 2-sided p-
`n (%)
`LSM of
`n (%)
`LSM of
`value)
`CFB2 (SE)
`CFB2 (SE)
`1.6 (0.6, 2.7; 0.0014)
`Clinical Summary Score3
`-3.0 (0.4)
`3643 (95)
`-4.6 (0.4)
`3638 (94)
`1.5 (0.5, 2.6; 0.005)
`-2.6 (0.4)
`3588 (94)
`-4.1 (0.4)
`3586 (93)
` Physical Limitation
`1.9 (0.8, 3.0; 0.001)
`-3.3 (0.4)
`3640 (95)
`-5.2 (0.4)
`3635 (94)
` Total Symptom
`2.2 (1.1, 3.3; <0.001)
`-3.0 (0.4)
`3637 (95)
`-5.2 (0.4)
`3632 (94)
` Symptom Frequency
`1.7 (0.6, 2.8; 0.003)
`-3.6 (0.4)
`3640 (95)
`-5.3 (0.4)
`3635 (94)
` Symptom Burden
`1LSM of difference = LSM of (CFB [LCZ696] - CFB [Enalapril]); 2CFB=change from baseline; 3Subject
`numbers represent subjects in the full analysis set with a KCCQ CSS score at both baseline and month 8 and
`subjects who died. An additional 171 subjects with data available at month 4 but not month 8 are included in the
`secondary endpoint analysis. 4One-sided p-value = 0.0007.
`Source: Table 23, Clinical Review
`
`In the prespecified analysis, subjects who died were assigned a Clinical Summary Score of
`zero for all subsequent visits. In the analysis shown below, subjects who died were treated as
`missing. The size of the treatment effect is smaller in this analysis (0.98; 95% CI 0.30, 1.66),
`suggesting that t