CENTER FOR DRUG-EVALUATION AND
`
`RESEARCH
`
`APPLICA TION NUMBER:
`
`22-341
`
`CROSS DISCIPLINE TEAM-LEADER REVIEW
`
`

`

`Cross Discipline Team Leader Review
`
`Cross-Discipline Team Leader Review - Addendum
`
`
`December 3, 2009
`
`Hylton V. Joffe, M.D., M.M.Sc.
`.
`
`Cross-Discipline Team Leader Review - Addendum
`22—341
`
`
`Novo Nordisk
`A . licant
`
`Date of Submission
`May 23, 2008
`PDUFA Goal Date
`March 23, 2009
`
`
` hi4;
`
`Proprietary Name /
`Victoza (liraglutide)
`
`
`Established (USAN) names
`Dosage forms / Strength
`6 mg/mL formulation administered subcutaneously via 3
`
`
`mL
`as 0.6 mg, 1.2 mg, or 1.8 mg
`Proposed Indication(s)
`As an adjunct to diet and exercise to improve glycemic
`
`
`control in adults with type 2 diabetes mellitus
`Approval, pending agreement on labeling.
`
`Recommended:
`
`Page 1 of 7
`
`

`

`Cross Discipline Team Leader Review
`
`This document is an addendum to the Cross Discipline Team Leader (CDTL) memorandum
`for Victoza (liraglutide). The purpose of this document is to summarize the following
`information that has become available afier the CDTL memorandum was finalized.
`
`1. A postmarketing case of liver failure in a liraglutide-treated patient
`2. Two postmarketing cases of gastric perforation in liraglutide-treated patients
`3. Updated calcitonin shift analyses, including follow-up information on the liraglutide-
`treated patient who had an increase in serum calcitonin to >50 ng/L as well as a
`postmarketing case of “suspected C-cell carcinoma”
`4. Recommendations from the recently completed immunology consult describing the
`need for better characterization of immunogenicity in a postmarketing clinical trial
`
`The information contained in this addendum does not alter my previous recommendation that
`liraglutide be approved.
`
`1. Postmarketing case of liver failure
`
`The sponsor submitted a safety report of liver failure regarding a 53~year old woman treated
`with liraglutide during an ongoing clinical trial. The patient was hospitalized with hepatic
`failure and encephalopathy after presenting with disorientation, Visual hallucinations, aphasia,
`and asterixis approximately 3 years after starting treatment with liraglutide. She had mildly
`elevated serum alanine aminotransferase (ALT), alkaline phosphatase, and gamma glutamyl
`transferase' (GGT) at screening and throughout the trial. Her ALT, alkaline phosphatase, and
`GGT values upon hospitalization were comparable to her screening and on-treatment values.
`Total bilirubin was 2.1 mg/dL (1.9x ULN). During the 3 years of liraglutide treatment, the
`patient’s total bilirubin values were mostly within normal limits, although on 5 occasions, her
`total bilirubin exceeded the upper limit of normal (maximum value prior to hospitalization was
`1.5x ULN at Week 92). The patient was treated with lactulose, improved within 72 hours and
`was discharged after a 9—day hospital stay. The cause for the liver failure is unknown — the
`patient did not have a history of alcohol abuse, was not taking any culprit medications, and
`tested negative for various causes of hepatitis, including hepatitis B and C, autoimmune
`hepatitis, alpha-l-antitrypsin, and Wilson’s disease. Of note, the transferrin saturation was
`56% suggesting hemochromatosis as a possible cause, although this condition was not further
`evaluated with genetic testing, there is no information on family history, and the patient was
`from Mexico (most cases of hemochromatosis occur in Caucasians). Liver biopsy confirmed
`cirrhosis but there is no mention of whether there was staining for iron. Liraglutide is not a
`likely explanation for the liver failure based on the fact that the patient had abnormal liver test
`measurements at screening that did not appreciably change during the treatment period. There
`is no signal for hepatotoxicity in the liraglutide new drug application, as discussed in the
`CDTL memorandum.
`
`Page 2 of 7
`
`'
`
`2
`
`

`

`Cross Discipline Team Leader Review
`
`2. Postmarketing cases of gastric perforation
`
`The sponsor submitted a 7-day safety report of a 52 year-old man in Germany who developed
`gastric perforation approximately two weeks after starting liraglutide. He was treated with
`laparatomy and oversewing of a gastric ulcer. The sponsor submitted another 7-day safety
`report of a 52 year-old man in Germany who developed gastric perforation and peritonitis
`approximately 1 week after starting liraglutide. The liraglutide dose at the time of both of these
`events was 1.2 mg. There are several similarities in the descriptions of these reports (same
`patient age, gender, country) and the sponsor is attempting to determine whether these 2
`reports pertain to the same patient. One of these reports describes the presence of an ulcer,
`which may havepredisposed the patient to gastric perforation. Liraglutide’s effects on
`delaying gastric emptying could conceivably cause greater distension/pressure in the stomach,
`leading to perforation in susceptible individuals. However, such a conclusion would be
`' premature based on limited information involving two (or possibly one) postmarketing cases.
`In addition, this has not been identified as a safety concern with the currently marketed
`glucagon—like peptide (GLP-l) agonist. The clinical reviewer for liraglutide should monitor for
`cases of gastric perforation post-approval via submitted 15-day Adverse Event Reporting
`System (AERS) cases and summary data in Periodic Update Safety Reports (PSURs).
`
`3. Calcitonin data
`
`The patient with an elevation in serum calcitonin to >50 ngzL:
`
`In the CDTL memorandum, there is mention of one liraglutide-treated patient (and no
`comparator—treated patients) who developed a treatment-emergent elevation in serum
`calcitonin to >50 ng/L. As mentioned in the CDTL memorandum, this 48 year-old man was
`treated with liraglutide 1.8 mg as add-on to glimepiride and had serum calcitonin values of
`10.7 ng/L at Week 0, 30.7 ng/L at Week 12 and 53.5 ng/L at Week 26. The patient did not
`report any thyroid-related adverse events. Afier finalization of the CDTL memorandum, we
`received updated information on this patient. He had a follow-up serum calcitonin of 22.3 ng/L
`obtained more than 2.5 years after the last dose of liraglutide with an estimated glomerular
`filtration rate of 56 mL/min suggesting mild renal impairment. The sponsor recommended that
`the patient be referred to an endocrinologist for further evaluation.
`
`The largest increase in serum calcitonin in a comparator-treated patient was seen with
`glimepiride in a patient whose serum calcitonin increased from 19.3 ng/L at baseline to 44.8
`ng/L at Week 65 and 38.1 ng/L at Week 104.
`
`7—day report of “suspected C-cell carcinoma”:
`
`On November 26, 2009, the sponsor submitted a 7-day report of “suspected C-cell carcinoma”
`in a patient receiving liraglutide 1.8 mg and insulin Detemir in an ongoing clinical trial. This
`diagnosis is based solely on serum calcitonin values — the patient has not yet undergone
`
`Page 3 of 7
`
`3
`
`

`

`Cross Discipline Team Leader Review
`
`thyroidectomy. Of note, this patient had a baseline (pre-liraglutide) serum calcitonin of 15.8
`pg/mL. Approximately 7 months later, while on liraglutide, the patient underwent pentagastrin
`stimulation testing. The peak serum calcitonin during this test was 128 pg/mL, prompting
`referral for thyroidectomy. However, the serum calcitonin was 16.1 pg/mL immediately prior
`to pentagastrin administration, which is similar to the baseline value of 15.8 pg/mL. Therefore,
`I agree with the sponsor’s assessment that the condition causing the calcitonin elevation was
`present prior to initiation of liraglutide and that it is unlikely that the exposure to liraglutide
`played a causal role1n the underlying thyroid abnormality.
`
`Updated shift data for serum calcitonin:
`
`Table 1 summarizes calcitonin shift data and is Virtually identical to Table 15 included in the
`CDTL memorandum. The only difference between Table 1 here and Table 15 in the CDTL
`memorandumIS the inclusion of data shown1n the shaded rows. These data were requested for
`completeness after the CDTL memorandum was finalized. The previously available data show
`the rates for patients meeting various calcitonin shift criteria over selected time periods (e.g..,
`first 20/24/26/28 weeks, first 52 weeks, 104 weeks) using last-observation-carried forward for
`missing data. The newly available data show the rates for patients meeting the calcitonin shift
`criteria at any point during their treatment with study medication.
`
`The bolded numbers in Table 1 correspond to rates that are numerically higher for liraglutide
`compared to the corresponding rates for placebo and active comparators. The liraglutide 0.6
`mg dose and the 1.2 mg dose were not more likely than comparators to meet the calcitonin
`shift criteria shown in Table 1. Each column in Table 1 contains 16 incidence rates and there
`are only 2/16 incidence rates for 0.6 mg and 1/ 16 incidence rates for 1.2 mg that are
`numerically higher for liraglutide compared to control. There are 10/ 16 incidence rates for 1.8
`mg that are numerically higher for liraglutide compared to control. However, it is noteworthy
`that the patients with the longest exposure to liraglutide (e.g., Week 104 data) did not have
`higher rates of calcitonin shifts compared to control. Lastly, a majority of the incidence rates
`(9/16) for total liraglutide were numerically lower than the corresponding incidence rates for
`control.
`
`The one liraglutide-treated patient with an increase in serum calcitonin to 250 ng/L is
`discussed above and the patients with an increase in serum calcitonin to 220 ng/L are
`discussed in the original CDTL memorandum. Note that the CDTL memorandum counts 11
`patients with a serum calcitonin increase to 220 ng/L when in fact there were 12 such patients
`(the twelfth patient is the patient described above who had an increase to 250 ng/L). This is
`clarified in the text below:
`
`A total of 11 liraglutide-treated patients (two with 0.6 mg, one with 1.2 mg, and eight with 1.8
`mg), five active comparator-treated patients, and one placebo-treated patient developed at least
`one treatment-emergent serum calcitonin 220 ng/L and <50 ng/L. One additional liraglutide-
`treated patient had an increase in serum calcitonin to 250 ng/L and is discussed separately
`above. One of the 11 liraglutide-treated patients with an increase in serum calcitonin to 220
`ng/L and <50 ng/L had an increase in serum calcitonin from 2.1 ng/L at baseline to 22.4 ng/L
`at Week 12. There are no additional calcitonin data becausethe patient was discontinued
`
`Page 4 of 7
`
`'
`
`~
`
`4
`
`

`

`Cross Discipline Team Leader Review
`
`prematurely due to nausea and diarrhea. For the remaining 10 liraglutide-treated patients with
`an increase in serum calcitonin to 220 ng/L and <50 ng/L, four had serum calcitonin values
`<20 ng/L at the last clinic visit despite continued treatment with liraglutide and the other six
`had increases in serum calcitonin from baseline to endpoint of only 2.1-7.1 ng/L with serum
`calcitonin at endpoint ranging from 20.2-25.8 ng/L (the patient with the 7.1 ng/L increase to
`25.8 ng/L at Week 26 was diagnosed with Hashimoto’s thyroiditis based on positive anti-TPO
`antibodies but had a normal thyroid ultrasound). For the six comparator-treated patients, four
`had serum calcitonin 220 ng/L at the last clinic Visit, with endpoint values ranging from 20.2-
`38.1 ng/L. In summary, eight liraglutide-treated patients (one on 0.6 mg, one on 1.2 mg, and
`six on 1.8 mg) and four comparator-treated patients had treatment-emergent serum calcitonin
`values 220 ng/dL at the last clinic visit, which is consistent with the overall patient—year
`exposures to liraglutide and control.
`
`In summary, the available shift data for serum calcitonin do not provide evidence of a
`convincing relationship to treatment assignment.
`
`APPEARS THlS WAY
`0N ORlGiML
`
`Page 5 of 7
`
`

`

`Cross Discipline Team Leader Review
`
`
`
`Shift from baseline
`
` N (safety dataset)
`
`Table 1. Calcitonin shift analyses using last observation carried forward
`u dates Table 15 in the CDTL memorandum
`————
`
`.
`.
`E
`.
`Active
`Total
`E
`
`'
`
`
` oooo-
`
`E
`
`=
`E
`
`1412
`'2122
`630
`846
`320
`, 381
`31511.1)
`,
`‘ 131 ]
`E15111)
`21(10)
`23.7
`22.7
`6(1.0)
`' 6(0.7)
`10.9
`8.1
`
`E
`
`210.5)
`2(O.6)
`35. E 3:0
`
`7(05) .-
`61 if
`E 8(04)
`E 5(04)
`7.9
`8.7
`210.3)
`210.2)
`3.6
`
`1
`
`2(0.6)
`
`. 5140.4)
`4.4,
`310.2)
`4.7
`210.3)
`3.6
`
`110.3)
`18
`
`.
`
`E
`
`
`
`1
`
`E
`
`210.2)
`2.7
`
`1103)
`11.5 .
`
`
`
`Week 20/24/26/28
`Week 52
`Week 104
`<ULN to persistently ZULN2
`Per 1,000 patient-years (PY)
`Week 20/24/26/28, n (%)
`Per 1,000 PY
`
`Week 52, n (%)
`Per 1,000 PY
`
`Week 104, n (%)
`Per 1,000 PY
`From <ULN to 21.511 ULN
`
`"
`
`Per 1,000 patient-years (PY)
`Week 20/24/26/28, n (%)
`Per 1,000 PY
`
`Week 52, n (%)
`Per 1,000 PY
`
`Week 104, n (%)
`Per 1,000 PY
`
`From <20 ng/L to 220 ng/L
`Per 1,000 patient-years (PY)
`Week 20/24/26/28, n (%)
`Per 1,000 PY
`
`Week 52, n (%)
`Per 1,000 PY
`
`Week 104, n (%)
`Per 1,000 PY
`
`1455
`E
`991
`479
`E
`497
`.
`.
`328
`E
`327
`4611.3) ;
`10(10),1 27
`15.1
`10.9
`E1.1)(10)
`27(1.9) 46(13)‘
`229
`41.8
`29.4
`1
`3(0.6)
`6(1.3)
`9(0.5)
`6.9
`14.1
`
`1(0.3)
`
`
`
`if»?!
`10103)
`6.4
`‘
`5(0.3)
`3.2
`
`310.4)
`
`, 12(1)3),
`39
`10 (03)
`6.4
`410.2)
`2.5
`
`1(1)1)
`
`From <50 ng/L to 250 ng/L
`Week 20/24/26/28, n (%)
`Per 1,000 PY
`Weeks 52 and 104
`includes liraglutide doses >0.6 to <1.2 mg and >1. 8 mg (used1n the two Japanese trials and1n the obesity tria
`2all values (even if only one is available) afier baseline >ULN
`Week 20/24/26/28— phase 3 diabetes trials, 20-week obesity trial, 24—week Japanese trials, 26~week exenatide trial
`Week 52 — monotherapy trial and extensions for the add-on to metformin trial, Japanese trials and obesity trial
`Week 104 — monotherapy and add-on to metformin extensions
`
`p—n)
`
`Page 6 of 7
`
`ooo'o‘v
`
`

`

`Cross Discipline Team Leader Review
`
`4. lmmunogenicity
`
`In the clinical development program, some patients developed anti-liraglutide antibodies that
`cross-reacted with native glucagon-like peptide (GLP)-l (see CDTL memorandum fOr details).
`The sponsor has not yet developed an assay to assess whether these cross—reacting antibodies
`are neutralizing. Dr. Rosebraugh, Director, Office of Drug Evaluation II, requested an
`immunogenicity consultation from the Division of Pulmonary and Allergy Products (DPAP)
`based on these findings. DPAP noted that cross-reactivity to endogenous GLP-l carries a
`potential risk of inactivation of the native protein and antigen-antibody complex-mediated
`disease. Therefore, DPAP is recommending that the sponsor evaluate the rate of anti-
`liraglutide antibody formation and potentially related adverse events after long-term dosing
`with liraglutide. DPAP stated that this assessment could be performed in a subset of patients in
`the required postmarketing cardiovascular trial or as a separate trial. In addition to antibody
`titers, DPAP recommended that the immunogenicity assessment include ongoing screening for
`laboratory parameters and adverse events related to inactivation of the native protein and
`possible antibody complex-mediated disease (e.g. cutaneous and musculoskeletal
`manifestations, complement levels, hepatic transaminases, and renal function). See Dr. Brian
`Porter’s review for further details.
`
`5. Recommendations
`
`I uphold my previous recommendation that liraglutide can be approved. As per the DPAP
`recommendations, immunogenicity must be further assessed in the postmarketing setting. This
`requirement will be communicated to the sponsor and will be reflected in the postmarketing
`requirements section of the approval letter.
`
`Page 7 of7
`
`7
`
`

`

`Application
`Type/Number
`
`Submission
`Type/Number
`
`Submitter Name
`
`Product Name
`
`NBA—22341
`
`ORlG-1
`
`N‘OVO NORDISK
`lNC
`
`VICTOZA (LIRAGLUTIDE)
`
`This is a representation of an electronic record that was signed
`electronically and this page is the manifestation of the electronic
`signature.
`
`HYLTON V JOFFE
`
`12/03/2009
`
`MARY H PARKS
`
`12/03/2009
`
`l concur with Dr. Joffe's addendum
`
`

`

`Cross Discipline Team Leader Review
`
`Cross-Discipline Team Leader Review
`
`
`October 14, 2009
`
`Hylton V. Joffe, M.D., M.M.Sc.
`
`Cross-Discipline Team Leader Review
`Subject
`
`NBA #
`22-341
`
`Applicant
`Novo Nordisk
`
`Date of Submission
`May 23, 2008
`
`PDUFA Goal Date
`March 23, 2009
`
`
`
`
`Victoza (liraglutide)
`Proprietary Name /
`
`Established tUSAN) names
`6 mg/mL formulation administered subcutaneously via 3
`Dosage forms / Strength
`
`mL 'f— as 0.6 mg, 1.2 mg, or 1.8 mg
`As an adjunct to diet and exercise to improve glycemic
`control in adults with type 2 diabetes mellitus
`Approval, pending agreement on labeling and a
`satisfactory response to the outstanding information
`- request
`
`Proposed Indication(s)
`
`Recommended:
`
`m4)
`
`Page 1 of 63
`
`1
`
`

`

`Cross Discipline Team Leader Review
`
`Cross Discipline Team Leader Review
`
`1. Introduction
`
`Glucagon-like peptide (GLP)-1 is released from the gastrointestinal tract during meals and
`stimulates insulin release from the pancreatic beta-cell in a glucose-dependent manner. GLP-l
`also reduces hepatic glucose production (by lowering glucagon secretion from the pancreatic
`alpha-cell) and slows gastric emptying. Endogenous GLP—l has a short half-life (<2 minutes)
`due to rapid degradation by dipeptidyl peptidase (DPP)-4. Because patients with type 2
`diabetes have reduced GLP-l concentrations, GLP-1 receptor agonists resistant to DPP—4
`degradation have been developed for the treatment of type 2 diabetes. Currently, Byetta
`(exenatide) is the only FDA-approved GLP-1 receptor agonist. Novo Nordisk has
`subsequently submitted a new drug application (NDA) for liraglutide (proposed tradename
`Victoza), a new GLP-1 receptor agonist, that is the focus of this memorandum.
`
`2. Background
`
`GLP-1 receptor agonists are protein-based therapies that are typically administered via the
`subcutaneous route. Other routes of administration under investigation include intranasal and
`skin (patch). Byetta is administered twice daily by subcutaneous injection. Byetta is currently
`not indicated for use as monotherapy because there were inadequate efficacy and safety data
`for this setting at the time of approval. The monotherapy indication is currently under review
`and will likely be approved in the near future. An NDA for a once-weekly formulation of
`Byetta (Exenatide LAR) is also under review.
`
`Safety concerns with Byetta include:
`0 Postmarketing reports of acute pancreatitis, including the more severe hemorrhagic or
`necrotizing forms with deaths
`0 Worsened renal function sometimes requiring hemodialysis, that may be attributed to
`dehydration due to gastrointestinal side effects
`Increased incidence of hypoglycemia when used in combination with a sulfonylurea
`
`0
`
`In addition, high titers of anti—exenatide antibodies may reduce efficacy.
`
`In July 2008, the Division convened a public, 2-day advisory committee meeting to discuss
`cardiovascular assessment for drugs and biologics developed for the treatment of type 2
`diabetes. After considering the recommendations of the advisory panel and other data, the
`Division published a December 2008 Guidance for Industry entitled Diabetes Mellitus —
`Evaluating Cardiovascular Risk in New Antidiabetic Therapies to Treat Type 2 Diabetes. This
`guidance document recommends that sponsors of new pharmacologic therapies for type 2
`diabetes show that these treatments do not result in an unacceptable increase in cardiovascular
`
`risk. Of note, the liraglutide NDA
`for the treatment of type 2 diabetes
`were submitted to FDA prior to the July 2008 advisory committee meeting and prior to the
`WW
`
`Page 2 of 63
`
`.
`
`2
`
`

`

`Cross Discipline Team Leader Review
`
`December 2008 guidance. Nonetheless, FDA has requested that the sponsors for these .—
`products provide adequate evidence of cardiovascular safety in accordance with the guidance
`to support approvability. Therefore, cardiovascular safety was a major focus of the clinical and
`statistical reviews for liraglutide. Another major focus of the liraglutide review pertains to
`findings of benign and malignant thyroid C-cell tumors in rodents. Both of these issues were
`discussed at a public advisory committee meeting on April 2, 2009 and the tumor issue was
`discussed at a regulatory briefing on June 26, 2009.
`
`N4)
`
`3. CMC/Device
`
`Liraglutide is produced by recombinant DNA technology in the yeast Saccharomyces
`cerevisiae. The drug substance for liraglutide is the 7—37 peptide fragment of human GLP-l
`with two modifications: Substitution of lysine by arginine at position 34 and addition of a
`glutamic acid-spaced palmitic acid to the lysine residue at position 26.
`The drug product consists of’ ._.. liraglutide, ’—— disodium phosphate dihydrate /-.—¥———,
`
`—— propylene glycol ( -
`_
`7.), and "—- . phenol (
`j and is supplied in a multiple-dose prefilled pen-injector. Each pen-injector
`
`contains 3 mL of drug product at a concentration of 6 mg/mL. 1
`
`“4)
`
`%. Part-way through the review cycle, the sponsor requested
`
`to also market a pen that can be used to administer
`. The Office of Surveillance
`and Epidemiology recommends that only this newly proposed pen be marketed because having
`
`7
`. is unnecessary and may lead to confusion. The sponsor has subsequently agreed
`
`to market only" ‘
`i
`‘
`
`The sponsor states that the liraglutide pens are modified versions of the previously cleared
`
`=
`. for insulin. The Center for Devices and Radiological Health (CDRH)
`review by Sajjad Syed dated February 13, 2009, raised concerns that the changes to the
`P— may introduce confusion when liraglutide users work with the modified pen-injector,
`prompting a request for a Human Factors study to show safe and effective use ofthe
`liraglutide pen-injector by the intended users. The sponsor conducted a Human Factors study,
`the design and results of which were found to be acceptable by CDRH to support
`approvability.
`
`4’4)
`
`Based on the results of stability testing, the Chemistry/Manufacturing/Controls (CMC)
`reviewers recommend a shelf-life for the drug product of 24-months at 2—8 degrees Celsius and
`32 days at 28-32 degrees Celsius.
`‘
`
`The drug product is photo-labile. The CMC reviewers note that the pen-injector adequately
`protects the drug product from degradation due to light.
`
`CMC has determined that the application qualifies for a categorical exclusion from an
`environmental assessment report because the expected introduction concentration of the active
`moiety at the point of entry into the aquatic environment is less than 1 part per billion.
`
`' Page 3 of 63
`
`3
`
`

`

`Cross Discipline Team Leader Review
`
`The Office of Compliance issued an acceptable recommendation on the manufacturing
`facilities of the drug product.
`
`CMC deficiencies identified during the review have been adequately resolved. The CMC
`reviewers have determined that the drug product is acceptable and recommend approval of the
`NDA. Please see reviews by Drs. Joseph Leginus, Ali Al-Hakim, Suong Tran, and Christine
`Moore for further details.
`'
`
`4. Nonclinical Pharmacology/Toxicology
`
`The non—clinical pharmacology/toxicology reviewers recommend not approving this NDA
`based on findings of thyroid C-cell tumors in mice and rats during the 2-year lifetime exposure
`carcinogenicity studies. Please see reviews by Drs. Anthony Parola and Karen Davis—Bruno
`for details. The finalized tertiary review by Dr. Paul Brown is pending at this time. The
`reviewers have concluded that the human relevance of these tumors is unknown. In the
`original NDA submission, the sponsor proposed that liraglutide-induces calcitonin secretion
`and synthesis driving C-cell hyperplasia and C-cell tumor formation in rodents, but not in
`primates. Dr. Parola did not agree that the conducted mechanistic studies support this proposed
`mode—of-action in rodents. For example, he notes that the proposed mechanism would be
`expected to first result in physiological, diffuse C-cell hyperplasia that precedes focal C-cell
`hyperplasia, but this did not occur. Our Executive Carcinogenicity Assessment Committee
`(ECAC) concurred that the sponsor did not provide adequate data on the animal thyroid C-cell
`tumor findings to demonstrate that these findings are not relevant to humans as did 12 of 13
`panel members at the April public advisory committee meeting. At the advisory committee
`meeting and in a subsequent face-to-face meeting, the sponsor abandoned the above-described
`mode-of—action, and instead, based human relevance on the absence of liraglutide effects on
`thyroid C-cells in primates, including the absence of treatment-related medullary thyroid
`carcinoma in clinical studies and the fact that liraglutide did not increase plasma calcitonin or
`proliferative C-cell lesions in monkeys treated for up to 20 months.
`
`Liraglutide tested negative in a standard battery of genotoxicity studies. Table 1 summarizes
`the findings from the rodent 2-year carcinogenicity studies. There is a dose-related effect of
`liraglutide on thyroid C—cell tumors in both genders in rats and mice. In rats, thyroid C—cell
`adenomas and carcinomas occurred at low multiples of clinical exposure. In mice, thyroid C-
`cell adenomas occurred at IO-times the clinical exposure and carcinomas occurred in 2/76
`(3%) females at 45-times the clinical exposure. Despite the thyroid C—cell tumor findings,
`liraglutide did not reduce survival in these studies.
`
`As discussed by Dr. Davis-Bruno, monkeys dosed with liraglutide for up to 20 months at ~60
`times human exposure did not develOp proliferative C-cell lesions, However, Dr. Davis—Bruno
`recommends caution in interpreting these findings because monkey studies are not powered or
`designed to evaluate carcinogenicity and the duration of treatment was only 5% of the monkey
`lifespan, which may not mimic long-term use of liraglutide in humans. In addition, liraglutide
`was immunogenic in monkeys, but not in rodents.
`
`Page 4 of 63
`
`4
`
`

`

`Cross Discipline Team Leader Review
`
`With regard to the rodent C-cell tumor findings, Dr. Parola notes that there is no direct
`evidence showing GLP-l receptors on thyroid C-cells, but states that indirect evidence
`suggests co-localization in rodents (e. g., GLP-l regulates bone resorption in mice through a
`calcitonin-dependent mechanism) and humans. In one study1 using autoradiography with 125I-
`GLP-1(7—36)amide, there was binding in 12/12 normal thyroid samples from rats, 3/5 normal
`thyroid samples from mice, and 1/18 normal thyroid samples from humans (the authors were
`unable to determine whether the binding occurred on thyroid follicular cells or on thyroid C-
`cells). This study also reported binding in 5/18 medullary thyroid cancer samples from
`humans. The sponsor attempted to directly evaluate whether the GLP-1 receptor is expressed
`in C~cells from thyroid tissue in rodents, monkeys and humans, but Dr. Parola has concluded
`that the findings are equivocal because of methodological issues. A summary of an
`autoradiographic ligand binding study submitted part-way through the NDA review cycle
`states that GLP—1 receptor binding occurred on C-cells in thyroid from rats, but not humans. A
`report for this study was not submitted for review. Nonetheless, it appears that'rodent C-cell
`tumors are a pharmacologic class effect due to persistent GLP—1 receptor activation. As
`discussed by Dr. Parola, there is preliminary evidence that other long-acting GLP—l agonists
`(dosed less frequently than once-daily) have a similar propensity to cause rodent C-cell tumors
`like liraglutide whereas short-acting GLP-l agonists are less tumorigenic. For example,
`exenatide, which is dosed twice daily, caused C-cell adenomas in female rats (no-observed-
`adverse-effect level <5x) but did not cause C-cell tumors in male rats or in mice (one potential
`limitation of the Byetta carcinogenicity studies is that dosing was once-daily whereas Byetta is
`clinically dosed twice-daily). In contrast, continuous subcutaneous infusion of exenatide in
`mice caused focal C-cell hyperplasia and the exenatide LAR formulation, which is intended
`for once-weekly closing in humans, appears to have considerably more tumorigenic effects on
`the thyroid C-cell than Byetta in rats (tumorigenicity with the exenatide LAR formulation has
`not yet been evaluated in mice).
`
`As discussed by Dr. Parola, there are six other approved medications that cause C-cell tumors
`in rats. Five of these drugs do so in only one gender. The seventh does so in both genders but
`at a no-observed-adverse-effect level of 20—times clinical exposure. None of these drugs cause
`C—cell tumors in mice.
`
`Diffuse and focal C—cell hyperplasia and C-cell tumors are common in rats (incidence >1%)
`but rare in mice (incidence <1%). C-cell carcinoma is rare (incidence <1%) in both species. In
`rats, the incidence of diffuse C-cell hyperplasia, focal C-cell hyperplasia, and C-cell adenomas
`increases with age. This increase in C-cell mass results in an increase in calcitonin with age.
`Dr. Parola has concluded that plasma calcitonin is a biomarker for liraglutide-induced C-cell
`tumors in mice, but not in rats. This means that in rats, liraglutide did not lead to consistent
`increases in calcitonin beyond the increases seen with age (which are due to increased C-cell
`mass). Therefore, calcitonin could still be a biomarker for increased C-cell mass in humans,
`even if caused by liraglutide, assuming these C-cells remain adequately differentiated to
`continue expressing calcitonin.
`
`I Korner M, et al. GLP-1 receptor expression in human tumors and human normal tissues: potential for in vivo
`targeting. J Nucl Med. 2007; 48: 736-43.
`
`Page 5 of 63
`
`5
`
`

`

`Cross Discipline Team Leader Review
`
`Dr. Parola calculated a time course for the C—cell tumor findings in rodents integrating data
`from control and high-dose groups from toxicity studies, mechanistic studies, and the
`carcinogenicity studies. In high-dose groups, the earliest appearance of thyroid C—cell
`carcinoma occurred after 64 weeks oftreatment in mice (~60% of their lifespan) with
`liraglutide exposures 45-times the clinical exposure and after 86 weeks oftreatment in rats
`(~70% of their lifespan) with liraglutide exposures 8-times clinical exposure. In young adult
`rats treated with liraglutide at exposures 8-times clinical exposure, liraglutide increased the
`, incidence of focal C-cell hyperplasia after 43 weeks of treatment and increased the incidence
`of C-cell adenomas after 30 weeks of treatment. In mice, focal C—cell hyperplasia occurred
`after only 4 weeks oftreatment with liraglutide at 88-times the clinical exposure and C-cell
`adenomas occurred as early as 47 weeks of treatment with liraglutide exposures 45-times the
`clinical exposure. There are at least two important limitations ofthese time-course data. First,
`toxicity studies are not designed or powered to evaluate carcinogenicity. Therefore, the
`absence of tumors in animals that complete toxicity studies should be interpreted in this
`context. Second, animals in the carcinogenicity studies are not sacrificed until dosing has been
`completed. Therefore, findings of tumors at earlier timepoints were detected in animals that
`died early for unrelated reasons. It is not possible to know when C-cell tumors first appeared in
`those animals that developed tumors but did not die early.
`
`Liraglutide has been approved by the European Medicines Agency (EMEA). Dr. Parola has
`reviewed the EMEA-assessment of the non-clinical thyroid C-cell tumors and agrees with their
`assessment that the findings in rodents are caused by a non-genotoxic mechanism that is
`probably GLP-l receptor-mediated, but disagrees that there are sufficient data to conclude that
`the relevance to humans was adequately assessed. Dr. Parola’s conclusion is that the relevance
`to humans is unknown at the present time.
`
`Based on the above findings, Dr. Parola is recommending that the sponsor determine the
`mode-.of-action for these tumors and evaluate the human relevance based on this mode-of-
`action. Activating mutations in the rearranged during transfection (RET) proto-oncogene
`account for most cases of familial medullary thyroid cancer (the human form of C-cell
`- carcinoma) and account for approximately one-half of patients with sporadic medullary
`thyroid carcinoma. RET mutations have not been identified in rat strains susceptible to C-cell
`carcinoma. Dr. Parola recommends that the sponsor evaluate the effect of liraglutide on RET
`signaling in thyroid C-cells in rodents and determine whether liraglutide alters phosphorylation
`of RET residues involved in C-cell proliferation and transformation. Dr. Parola also
`recommends that the sponsor assess whether the thyroid GLP—1 receptor is required for
`liraglutide’s proliferative effects and whether liraglutide—induced C-cell tumors occur in GLP-
`1 receptor knockout animals or rodents treated with a GLP-1 receptor antagonist.
`
`Page 6 of63
`
`6
`
`

`

`
`
`
`
`:80;:80$3VN§meNNGas3$3333Naxe;ONGas3§NNV:3