CENTER FOR DRUG EVALUATION AND
`
`' RESEARCH
`
`APPLICATION NUMBER:
`
`22-350
`
`CROSS DISCIPLINE TEAM LEADER REVIEW
`
`

`

`Cross Discipline Team Leader Review
`
`Cross-Discipline Team Leader Review
`
`
` _
`July 28, 2009_
`
`
`From
`'
`Hylton V. Joffe, M.D., M.M.Sc.
`
`
`Subject
`Cross-Dis_cipline Team Leader Review
`
`
`-
`NBA #
`22-350
`
`
`
`
`
`Bristol-M ers Suibb
`Applicant
`‘
`
`
`June 30, 2008
`'
`Date of Submission
`
`
`
`July 30, 2009
`PDUFA Goal Date
`
`
`
`
`Proprietary Name /
`‘
`Onglyza (saxagliptin)
`Established (USAN) names
`
`
`
`
`
`Dosa e forms/ Strflgth
`2.5 mg and 5 mg tablets
`
`
`
`Proposed Indication(s)
`As an adjunct to diet and exercise to improve glycemic
`
`
`
`control in adults withtlm: 2 diabetes mellitus
`
`
` Recommended:
`
`Approval, _: ending agreement on labeling
`
`
`
`Page 1 of 53
`
`1
`
`

`

`Cross Discipline Team Leader Review
`
`Cross Discipline Team Leader Review
`
`1. Introduction
`
`. Incretin hormones, such as glucagon-like peptide (GLP)-1 and glucose—dependent
`insulinotropic polypeptide (GIP), are released from the gastrointestinal tract during meals and
`stimulate insulin release from the pancreatic beta-cell in a glucose-dependent manner. GLP-1
`and GIP have short half-lives (<2 minutes) due to rapid degradation by dipeptidyl peptidase
`(DPP)-4. Saxagliptin (proposed tradename Onglyza) is an oral DPP-4 inhibitor that has been
`develbped by Bristol-Myers Squibb as anadjunct to diet and exercise to improve glycemic
`control in adults with type 2 diabetes. The original user fee goal date for this application was
`April 30, 2009. An unexpected teratogenicity finding in a rat embryofetal development study
`designed to support the saxagliptin/metformin fixed-dose combination tablet (see below)
`prompted submission of the non--clinical study report to FDA within 3 months of the user fee
`goal date. The Division classified this submission as a major amendment and extended the user
`fee goal date by 3 months to July 30, 2009.
`
`This memorandum discusses the saxagliptin new drug appliCation (NDA) with a focus on key
`findings from the various review disciplines and the phase 2/3 development program.
`
`2. Background
`
`DPP—4 inhibitors tend to have modest efficacy but these medications appear to be generally
`well-tolerated with neutral effects on body weight and a low risk for hypoglycemia. Currently,
`Januvia(sitag1iptin phosphate) is the only FDA-approved DPP'—4 inhibitor. Labeled safety -
`concerns with Januvia include postmarketing reports of hypersensitivity reactions, including
`Stevens—Johnson Syndrome, and minor increases'in serum creatinine in patients with moderate
`or severe renal impairment. Postmarketing reports of pancreatitis in association with Byetta
`and Januvia are under FDA review. Other toxicities associated with at least one DPP-4
`inhibitor include necrotic skin lesions in monkeys, sometimes near clinical exposures (e. g.
`vildagliptin, dutogliptin) and possible hepatotoxicity (vildagliptin).
`
`» 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 committee panel and other
`data, the Division published a December 2008 Guidance for Industry entitled Diabetes
`Mellitus— Evaluating Cardiovascular Riskin New Aniidiabetic Therapies to Treat Type 2
`Diabetes. This guidance document requests that sponsors ofnew pharmacologic therapies for
`type 2 diabetes show that these treatments do not resultin an unacceptable1ncrease in
`cardiovascular risk. Of note, the saxagliptin NDA and two other NDAs for the treatment of
`type 2 diabetes were submitted to FDA prior to the July 2008 advisory committee meeting and
`prior to the December 2008 guidance. Nonetheless, FDA has requested that the sponsors for
`these three products provide adequate evidence of cardiovascular safety in accordance with the
`
`Page 2 of 53
`
`2
`
`

`

`Cross Discipline Team Leader Review
`
`guidance to support approvability. Therefore, cardiovascular safety was a major focus of the
`clinical and statistical reviews for saxagliptin.
`
`3. CMC
`
`The chemistry/manufacturing/controls (CMC) portion of the NDA was submitted as part of the
`Office of New Drug Quality Assessment (ONDQA) Quality—by--Design Pilot Program to
`explore science and risk—based approaches to assuring product quality. The drug substance for
`OnglyzalS saxagliptin C
`)availablem dOsage strengths of 2. 5 mg and 5 mg. The
`saxagliptin molecule contains chiral centers but there is no chiral conversion in vivo. The drug
`
`h“)
`
`)
`productdoesnotcontainnovel excipientsandismanufacturedusinga ‘
`Based on stability data, the CMC
`reviewers are granting the two dosage strength presentations a 36-month stability period with
`labeling that states “Store at 25°C (77°F); excursions permitted to 15°~30°C (59°-86°F) [see
`USP Controlled ROOm Temperature].”
`
`f K
`
`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.
`
`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. All Drug
`Master Files are acceptable or the pertinent information has been adequately provided.
`The CMC reviewers have determined that the drug product is acceptable and recommend
`approval of the NDA. Please see reviews by Drs. Sharrnista Chatterjee, John Hill, Prafull
`Shiromani, and Christine Moore for further details.
`I
`
`4. Nonclinical Pharmacology/Toxicology
`
`The pharmacology/toxicology reviewers have concluded that there are reasonable safety
`margins between animal toxicities and clinical exposures with the proposed maximum daily
`dose of 5 mg, and recommend approval pending agreement on labeling. As explained below,
`the reviewers are also recommending two required non-clinical postmarketing studies to
`further explore an unexpected teratogenicity finding in a rat embryofetal development study
`that co-administered saxagliptin and metformin. The sponsor has already initiated these studies
`but the pharmacology/toxicology reviewers have determined that the protocols are inadequate
`(e.g., too high a dose of metforrnin is being tested). The Division is in the process of
`communicating the protocol inadequacies tothe sponsor and has informed the sponsor that
`new studies will be needed. Please see reviews by Drs. Fred Alavi, Todd Bourcier, and Paul
`Brown for further details.
`
`Page 3 of 53
`
`3
`
`

`

`31(4)
`
`Cross Discipline Team Leader Review
`
`In this embryofetal study, tvsio fetuses froni one
`,
`i
`L - A
`litter of rats that had been exposed to co-administered saxagliptin (at doses 114—fold above
`clinical exposures) and metformin (at doses 4-fold above clinical exposures with the 2,000 mg
`. daily dose) developed malformatiOns (one case of craniorachischisis, a rare neural tube defect
`involving incomplete closure of the skull and spinal cord, and one case of cleft palate). The
`sponsor attributed this finding to metformin alone (via potential alteratidns to vitamin Biz and
`folate) but the embryofetal study did not have a metformin alone treatment arm to support this
`assertion. Furthermore, Dr. Bourcier notes that this study had two combination dose groups
`that had equal exposures to metformin, yet teratogenicity occurred only in the group with the
`higher dose of saxagliptin. In addition, Dr. Bouricer notes that the original embryofetal studies
`conducted for metformin did not report craniorachischisis.
`
`Based on this finding, the pharmacology/toxicology reviewers are recommending that the
`Division require two non—clinical postmarketing studies under the FDA Amendments Act
`(FDAAA), one in rats and another in rabbits, that further explores this signal using study
`designs that include separate treatments arms for metformin alone, saxagliptin alone, and the
`combination of saxagliptin+metformin.
`
`,
`
`Results from this study have relevance to the saxagliptin NDA because saxagliptin will be
`frequently co-administered with metformin, if approved. Therefore, the pharmacology/
`toxicology reviewers are recommending a statement in the label describing this finding and
`will reassess the labeling once results from these two more definitive studies are available. Of
`note,_ the reviewers (with input from the Associate Director and Director of ‘
`PharmacOlogy/Toxicology and the Reproductive Toxicology Subcommittee at FDA) are
`recommending Pregnancy Category B for saxagliptin, because saxagliptin alone was not
`teratogenic in rats and rabbits at very high exposure multiples.
`
`Saxagliptin is metabolized by CYP3A4 to active metabolite EMS-510849 in all test species.
`This active metabolite is two-fold less potent than saxagliptin but more selective at inhibiting
`DPP-4 versus off-target DPP-8 and DPP-9. In patients with diabetes, exposures to this
`metabolite are 4-7-fold higher than exposures to saxagliptin whereas in animals, exposures to
`this metabolite are no greater than exposures to saxagliptin. Nonetheless, Dr. Alavi has
`determined that this metabolite and several minor metabolites formed in humans have been
`
`adequately assessed for toxicity in the non-clinical studies.
`
`Brain lesions occurred in male rats administered very high doses (>350-fold safety margin) of
`saxagliptin. Per Dr. Bourcier, the sponsor has convincingly demonstrated that these lesions are
`caused by release of cyanide from saxagliptin via CYP2C11, an androgen-regulated
`metabolizing enzyme that is abundant in male rats but not present in humans. Per Dr. Alavi,
`saxagliptin administration to female» rats and to both genders of other species used in non-
`clinical studies did not lead to measurable quantities of cyanide. In addition, whole blood
`.
`cyanide concentrations were below the limit of quantification in all healthy volunteers
`receiving up to 40 mg of saxagliptin daily for 14 consecutive days in Study 181031. Dr. Alavi
`
`Page 4 of 53
`
`.
`
`4
`
`

`

`Cross Discipline Team Leader Review
`
`has determined that there is little 0r no risk that these findings in male rats are applicable to
`humans.
`
`Some DPP-4 inhibitors cause cutaneous lesions in monkeys, possibly due to cross-reactivity
`with DPP-8 and DPP-9. Sitagliptin, the only FDA-approved DPP-4 inhibitor, does not cause
`these lesions but saxagliptin does. However, per Dr. Alavi the risk of skin lesions to humans
`treated with saxagliptin is minimal because there are large safety margins for this toxicity
`(exposures 20-fold above the clinical dose cause only minimal non-neCrotizing cutaneous
`lesions; exposures approximately 60-fold above the clinical dose cause severe necrotizing
`lesions). Saxagliptin also induced minimal erosive lesions of the paws in dogs after 12 months
`exposure at doses BBS-fold above clinical exposure.
`
`Dr. Bourcier notes that splenic lymphoid proliferation and multi-organ lymphoid/monocytic
`infiltration occurred in all animal species but characterizes these findings as minimal and
`reversible. Reductions in lymphocyte counts occurred in rats, dogs, and monkeys given high
`doses of saxagliptin. However, Dr. Bourcier states that the animal data do not provide much
`insight into the reduction in lymphocytes reported in humans (see the safety section of this
`memorandum) because the animal finding was not consistent across studies and did not follow
`dose- or time-dependency. Dr. Bourcier concludes that there is minimal to no risk of severe
`immunotoxicity but cannot exclude subtle changes in immunity.
`
`Exposure in human milk is expected because saxagliptin is detected in rat milk.
`
`In the 2-year rat and mouse carcinogenicity studies, there were no drug-related increases in
`neoplastic lesions despite high exposures to study drug (up to 1,000— to 2,200-fold margins for
`saxagliptin and up to 68- to 300—fold margins for active metabolite EMS-510849).
`
`There is no evidence of cardiovascular toxicity based on non-clinical testing in healthy
`animals.
`
`5. Clinical Pharmacology/Biopharmaceutics _
`
`The Clinical Pharmacology reviewers recommend approval pending agreement on labeling.
`Please see the joint review co—authored by Drs. Jayabharathi Vaidyanathan, Immo Zdrojewski,
`and Justin Earp for details.
`'
`
`As mentioned above, saxagliptin has one major metabolite (EMS—510849) that is also an
`inhibitor of DPP-4. This metabolite is two-fold less potent than saxagliptin but has greater
`selectivity than saxagliptin for DPP-4 over DPP-8 (948—fold vs. 391-fold) and DPP—9 (163-fold
`vs. 75—fold) at 37°C.
`'
`
`Saxagliptin and EMS—510849 have negligible protein binding. The median Tmax for
`saxagliptin is between -1 5—2.0 hours and the elimination half—life is 2.3-3.3 hours. BMS—
`510849 has a median Tmax of3.0 hours and a mean apparent terminal half-life of 3.6 hours.
`Saxagliptin has similar pharmacokinetics in patients with type 2 diabetes and healthy subjects.
`However, in patients with type 2 diabetes mean exposures to B'MS-S 10849 are 4—7—fold higher
`
`Page 5 of 53
`
`-
`
`.
`
`5
`
`

`

`Cross Discipline Team Leader Review
`
`than exposures to saxagliptin, whereas healthy subjects have mean exposures to BMS-S 10849
`that are 2-3 fold higher than exposures to saxagliptin.
`
`The amount of DPP-4 inhibition at 24 hours is 37% with the 2.5 mg saxagliptin dose and 65%
`with the 5 mg saxagliptin dose.
`.
`.
`:
`
`Saxagliptin can be dosed regardless of food. No dosage adjustment is needed based on age,
`race, or gender. Table 1 summarizes the percent changes in mean exposures to saxagliptin and
`EMS-510849 in patients with various degrees of renalimpairment.
`
`
`
`BMS—510849
`
`Table 1. Percent changes in mean exposures to saxagliptin and
`metabolite EMS-510849 in patients with renal im - airment
`
`Degree of renal
`Saxa li tin
`
`
`
`
`impairment
`AUC0_T
`Cmax
`
`
`
`
`Mild '
`+15% l +39%
`
`
`
`
`+191% (29-fold)
`+47% '
`Moderate
`+40%
`+7%
`
`
`
`+347% (45-fold) _
`+46%
`Severe
`+110%
`+3 8%
`
`
`
`
`
`
`+306% 4.1—fold
`+36%
`Hemodial sis
`-23% j
`-15% I
`
`
`AUC = area under the time—concentration curve
`
`
`
`
`Based on these data, the clinical pharmacology reviewers agree with the sponsor’s proposed
`dosage adjustment to 2.5 mg for moderate, severe and end-stage renal impairment and agree
`that no dosage adjustment is needed for mild renal impairment. Dr. Bourcier notes that the
`moderate increase in drug exposure in patients with renal impairment who inadvertently
`receive the unadjusted 5 mg clinical dose is unlikely to reproduce toxicities noted in animals.
`
`Sax'agliptin is predominantly metabolized in the liver by CYP3A4/5 and is a P-glycoprotein
`substrate. In the hepatic impairment pharmacokinetic study, subjects with severe hepatic
`impairment had a geometric mean decrease in saxagliptin Cmax of 6% and a geometric mean -
`increase in overall saxagliptin exposure (area under the time—concentration curve or AUC) of
`approximately 75% compared to subjects with normal hepatic impairment. Based on these
`data, the clinical pharmacology reviewers agree with the sponsor that no dosage adjustment is
`needed in‘ patients with hepatic impairment.
`
`Drug interaction studies were conducted with ketoconazole, diltiazem, rifampin, Maalox Max,
`famotidine, omeprazole, glyburide, pioglitazone, metformin, digoxin and simvastatin. The
`spOnsor did not conduct a drug interaction study with warfarin. The oral contraceptive drug
`interaction study has not yet been completed.
`'
`‘
`
`CYP3A4/5 induction with rifampin reduced saxagliptin AUC by 80% and increased BMS—
`510849 Cmax by 40% with a five-fold increase in the metabolite-to—parent AUC ratio.
`CYP3A4/5 induction also reduced the saxagliptin half-life from 3.0 hours to 1.7 hours. Dr.
`. Vaidyanathan states that the clinical significance of these changes is unknown but is not
`recommending dosage adjustment because there is only a 25% decrease in exposure to the
`total active moiety (molar parent exposure + one-half molar metabolite exposure).
`
`Page 6 of 53
`
`‘
`
`-
`
`6
`
`

`

`Cross Discipline Team Leader Review
`
`The sponsor conducted two drug interaction studies with ketoconazole, which is a strong
`CYP3A4/5 inhibitor. In the first ketoconazole study (CV181005), subjects received saxagliptin
`100 mg on Days 1 and 9 and ketoconazole on Days 3—1 1. Ketoconazole resulted in a 2.5-fold
`geometric mean increase in saxagliptin AUC. All subjects had normal lymphocyte counts at
`baseline. On Day 10, fourteen of the 15 subjects had a reduction from baseline in lymphocyte
`counts (range -14% to -80%). Seven of these subjects had Day 10 lymphocyte counts below
`the lower limit of the reference range that returned to within the reference range on Day 12.
`Five of these seven subjects with lymphopenia developed fever and chills in the evening on
`Day 9 and all seven patients had reductions from baseline in platelet counts on Day '10 (range -
`7% to -24%), although the platelet counts remained within the reference range.
`
`In the second ketoconazole study (CV181022), subjects were randomized to Sequence 1
`(saxagliptin 5 mg on Days '1- and 9), Sequence 2 (saxagliptin 20 mg on Days 1 and 9),.or
`Sequence 3 (saxagliptin 20 mg on Days 1 and 9 and ketoconazole on Days 3—11).
`Ketoconazole resulted in a 3.8-fold geometric mean increase in saxagliptin AUC. There were
`no reports of pyrexia or chills in the subjects randomized to Sequence 3 but these subjects had
`a 31% mean reduction in absolute lymphocyte counts compared to” a mean reduction of 22%
`for Sequence 1 and a mean reduction of 19% for Sequence 2.
`
`It is unknown to what extent ketoconazole will increase exposures for a 5 mg saxagliptin dose.
`However, based on the 3.8-fold geometric mean increase in saxagliptin AUC with
`ketoconazole in Study CV181022, the clinical pharmacology reviewers are recommending a
`dose reduction to 2.5 mg for patients co-administered strong CYP3A4/5 inhibitors.
`
`;
`
`The sponsor also conducted a drug interaction study with diltiazem, which is a moderate
`' CYP3A4/5 inhibitor. Subjects received saxagliptin 10 mg on Days 1 and 9 and diltiazem on
`Days 2-10. None of the subjects reported fever or chills. Three of the 14 subjects had
`lymphocyte counts at study end that were lower than the lymphocyte counts at screening or
`baseline, but these 3 subjects hadlymphocyte counts well within the reference range.
`Diltiazem resulted in a 2.1-fold geometric mean increase in saxagliptin AUC. Based on these
`results, the clinical pharmacology reviewers are not recommending saxagliptin dosage
`adjustment for patients on moderate CYP3A4/5 inhibitors.
`
`'
`
`The clinical pharmacology reviewers have concluded that none of the other tested drug
`interactions are likely to be clinically relevant.
`
`Saxagliptin has no significant prolongation effect on the QT interval. In the Thorough QT
`Study, 40 healthy subjects were randomized in a double-blind, crossover fashion to each of the
`following four treatments: saxagliptin 10 mg, saxagliptin 40 mg (8—fold higher than the
`proposed 5 mg maximum recommended dose), placebo, and moxifloxacin 400 mg (positive
`control). Moxifloxacin was given as a single dose Whereas the other treatments were given
`once daily for 4 days. Per the QT Interdisciplinary Review Team, the upper limits of the two—
`sided 90%”confidence intervals for the mean QT difference between saxagliptin and placebo
`were below 10 msec, the threshold for regulatory concern as described in the International
`Conference on Harmonisation (ICH) E14 guideline. Please see Dr. Christine Garnett’s review
`under the saxagliptin investigational new drug application (IND) for further details.
`‘
`
`Page 7 of 53
`
`7
`
`

`

`Cross Discipline Team Leader Review
`
`A new tablet color and embossing are the only differences between the phase 3 formulations
`and the to—be—marketed formulations. At the pre-NDA meeting, the Division agreed that there
`is no need for a pivotal bioequivalence study based on these minor differences.
`
`The clinical pharmacometn'cs group modeled the relationship between saxagliptin exposure
`and absolute lymphocyte count using phase 3 monotherapy study CV18101 1 and population
`pharmacokinetic data. The reviewers concluded that the decrease in absolute lymphocyte count
`is linear to the increase of the total active moiety exposures within the tested daily saxagliptin
`dose range of 25-10 mg. The mean reduction in lymphocyte count with 5 mg and 10 mg was
`4% at 24 weeks. Please see the safety section of this memorandum for further details.
`
`6. Clinical Microbiology
`
`Not applicable.
`
`7. Clinical/Statistical- Efficacy
`
`This section will focus on the efficacy results from the controlled, phase 2/3 clinical trials,
`' which consists of one 12—week phase 2 dose-ranging study and six 24—week phase 3 clinical
`trials. Please see Dr. Naomi Lowy’s clinical'review for fiirther details.
`
`The NDA includes a 12-week trial (CV181041) evaluating saxagliptin 5 mg vs. placebo on
`measuresof beta-cell function in treatment-naive patients with type 2 diabetes. This trial did
`not have, primary or secondary traditional efficacy endpoints (e. g. HbAlc) and only 36 patients
`were randomized. Therefore, this trial will not be discussed in this memorandum even though
`the sponsor classified it as a phase 3 study.
`
`The phase 2 dose-ranging study (CV181008) compared the efficacy and safety of saxagliptin
`2.5- mg, 5 mg, 10 mg, 20 mg, and 40 mg once daily to placebo in patients with inadequate
`glycemic control on diet and exercise. The protocol was amended 6 months into the trial to
`allow for randomization of additional patients to a 100 mg dose group of saxagliptin vs. a
`second placebo group for 6 weeks.
`
`The phase 2/3 clinical trials included in the original NDA were randomized, multinational,
`double-blind, and either placebo- or active-controlled. The phase 3 trials had a l-4-week
`placebo run—in period prior to randomization. Saxagliptin was to be taken prior to the morning
`meal except for one of the treatment arms in monotherapy study CV181038 (see below),
`which also evaluated the effect of dosing saxagliptin prior to the evening meal.
`
`The 24—week phase 3 clinical trials evaluated saxagliptin in the following settings:
`
`CV181011 and CV181038: Monotherapy trials in treatment-naive patients
`0 Patients were to have 53 consecutive days and <7 non-consecutive days of anti-diabetic
`therapy within the 8 weeks prior to screening, and <6 months total of prior anti-diabetic
`therapy.
`
`Page 8 of 53
`
`V
`
`8
`
`

`

`Cros’s Discipline Team Leader Review
`
`0
`
`0
`
`Study -01 1 compared saxagliptin 2.5 mg, 5 mg, and 10 mg vs. placebo. This study also
`included an open-label, non—randomized cohort with screening HbAlc >lO% to 12% who
`were treated with saxagliptin 10 mg daily. This uncontrolled cohort will not be extensively
`A reviewed1n this memorandum.
`
`Study -038 compared saxagliptin 2.5 mg AM (dosed prior to the morning meal), 5 mg AM
`5 mg PM (dosed prior to the evening meal), and 2. 5 mg AM with possible titration to 5 mg
`AM (based on prespecified fasting plasma or Whole blood glucose values at Weeks 4, 8,
`12, and 24) vs. placebo. The primary objective compared the AM treatment arms to
`placebo. A secondary objective compared the PM treatment arm to placebo. The protocol -
`did not prespecify a comparison between the AM and PM dosing regimens.
`
`CV181013: Add-on to thiazolidinedione
`0 Patients were to be taking rosiglitazone 34 mg/day or pioglitazone 230 mg/day
`monotherapy for 212 weeks prior to screening.
`.0 This trial compared add-on saxagliptin 2.5 mg and 5 mg vs. add-on placebo.
`
`CV181014: Add-on to metformin
`
`-
`
`Patients were to be taking stable metformin monotherapy (1,500—2,550 mg) for >8 weeks
`prior to screening.
`0 This trial compared add—on saxagliptin 2.5 mg, 5 mg, and 10 mg vs. add—on placebo.
`
`_
`CV181040: Add-on to sulfonylurea
`0
`Patients were to be taking a submaximal dose (less than the maximum approved dose) of
`sulfonylurea monotherapy for >2 months prior to screening.
`0 After the 4-week run-in period, patients discontinued their current sulfonylurea therapy and
`started open-label glyburide 7.5 mg daily.
`0 This trial compared add-on saxagliptin 2.5 mg and 5 mg vs. add-on placebo + uptitrated
`glyburide. The add-on placebo + uptitrated glyburide treatment arm received placebo plus
`2.5 mg of blinded glyburide as add-on to the 7.5 mg background dose of open-label _
`’
`glyburide. In this treatment arm, the blinded glyburide dose was uptitrated to a maximum
`dose 0f 7.5 mg (total daily glyburide dose of 15 mg) at Weeks 2 and 4 based on
`prespecified criteria for fasting plasma or whole blood glucose.
`
`CV181039: Initial combination with metformin
`
`0
`
`Patients were to have S3 consecutive days and <7 non-consecutive days of anti-diabetic
`therapy within the 8 weeks prior to screening, and <1 month total of prior anti-diabetic
`therapy.
`
`0 This trial compared saxagliptin 10 mg + metformin, saxagliptin 10 mg + placebo,
`metformin + placebo, and saxagliptin 5 mg + metformin.
`0 The 3, treatment arms randomized to metformin started 500 mg of the immediate—release
`formulation that was blindly uptitrated through Week 5 to a maximum of 2,000 mg/day in
`divided doses based on prespecified criteria for fasting plasma or whole blood glucose.
`0 This trial did not include a saxagliptin 5 mg + placebo treatment arm, limiting the ability to
`assess the contribution of saxagliptin 5 mg to the saxagliptin 5 mg + metformin arm.
`
`Page 9 of 53
`
`.
`
`I
`
`9
`
`

`

`Cross Discipline Team Leader Review
`
`These clinical trials had similar inclusion and exclusion criteria. Entry criteria included age
`between 18—77 years and type 2 diabetes with baseline HbAlc 7-10% (75-10% for the add—on
`to sulfonylurea trial; 7-10.5% for the add-on to thiazolidinedione trial; 8-12% for the initial
`combination with metformin trial). Exclusion criteria included elevated serum creatinine (21.5
`mg/dL for men and 21.4 mg/dL for women in the monotherapy trials and in the trials that used
`metformin as background or randomized therapy; 22.0 mg/dL for the add-on to sulfonylurea
`and add-on to thiazolidinedione trials) and history of significant cardiovascular history, such as
`myocardial infarction, coronary intervention, or cerebrovascular accident within the 6 months
`prior to study entry. These trials also excluded patients with a history of New York Heart
`Association Class III or IV heart failure or known left ventricular ejection fraction 540%.
`
`The primary efficacy timepoint was 24 weeks for all phase 3 trials. The sponsor labeled the
`24-week period of each trial as the “short-term” phase. Each phase 3 trial had a “long-term?
`phase of at least 12 months duration that followed the short—term phase. Patients continued
`receiving double—blind study medication during the long—term phase. Patients were eligible to
`enter the long-term phase upon completion of the 24-week short-term phase or upon early
`discontinuation from the short-term phase because of the need for glycemic rescue therapy.
`
`. Three of the phase 3 trials had treatment regimens in the long-term phase that differed from the
`treatments in the corresponding short-term phase:
`.
`In monotherapy studies ~011 and -03 8, the placebo arm in the short-term phase was re-
`assigned to placebo + blinded metformin 500 mg daily in the long-term phase
`In monotherapy study -03 8, all saxagliptin groups could be titrated to 10 mg according to
`prespecified HbAlc criteria in the long-term phase
`7
`o ‘ In the add-on to sulfonylurea trial, glyburide in the placebo + uptitrated glyburide arm
`could be titrated to 20 mg (vs. titration to 15 mg in the short-term phase)
`
`0
`
`All short—term treatment periods were completed at the time of NBA submission whereas
`interim data were presented for the long—term treatment periods. Ms. Joy Mele, the biostatistics
`reviewer, focused the efficacy evaluation on the data from the short—term phase only. I agree
`with this approach because the efficacy data from the long-term phase are confounded by
`glycemic reScue therapy, are based on only a subset ofrandomized patients (79-90%) entering
`the long—term phase, and are limited by smaller sample sizes at the time of the long-term
`efficacy assessments (24-72% of patients discontinued from the combined short-term and
`long-term phases as ofthe l20-day safety update).
`
`As shown in Tables 5.4 and 5.5 in Dr. Lowy’s review, glycemic rescue criteria were similar,
`but not identical, across the phase 3 trials. For the 24-week short-term phase, progressively
`more stringent cutpoints for fasting plasma or whole blood glucose were used to prompt
`initiation of glycemic rescue therapy. In the long—term phase, progressively more stringent
`cutpoints fer HbAl c were used to prompt initiation of glycemic rescue therapy, permitting
`HbAlc measurements as high as 8.0% early during the long—term phase but only allowing
`HbAlc as high as 7.0% towards the latter part ofthe long-term phase. The open-label rescue
`therapy was metformin for‘all phase 3 trials (initiated at 500 mg and titratable to 2,000 or
`2,500 mg) except for the add-on to metformin trial and the initial combination with metformin
`trial, which used pioglitazone as rescue (initiated at 15 mg and titratable to 30 or 45 mg).
`
`Page 10 01°53
`
`I
`
`10
`
`

`

`Cross Discipline Team Leader Review
`
`The objective'of all phase 3 trials was to Show superiority of saxagliptin over control on the
`primary efficacy endpoint of change from baseline to Week 24 in HbAlc. Other efficacy
`endpoints included change from baseline in fasting plasma glucose (FPG), HbAlc responder
`analyses, change from baseline in AUC from 0-180 minutes for post-prandial glucose response
`to an oral glucose tolerance test, and the proportion of patients requiring glycemic rescue,
`failing to achieve pre-specified glycemic targets, or discontinuing for lack of efficacy. The
`AUC endpoint for post-prandial glucose is not readily interpretable to clinicians. The -
`~ remaining endpoints are typical for trials designed to support approvability of anti—diabetic
`medications.
`
`As discussed by Ms. Mele the primary statistical population for each trial iconsisted of all
`randomized patients with a baseline and at least one post-baseline assessment of the parameter
`of interest. The last-observation-carried—forward (LOCF) method was used for patients with
`missing data and for patients who initiated glycemic rescue therapy, The primary efficacy
`analysis was conducted using analysis of covariance (ANCOVA) with baseline HbAlc as a
`covariate.
`'
`'
`
`As noted by Dr. Lowy, the Russian government suspended the export of biological samples for
`several weeks, requiring the sponsor to use an emergency central laboratory in Moscow for all
`Russian sites involved in the conduct of monotherapy study —038 and the initial combination
`with metformin trial. This government suspension affected samples for HbAlc and glucose,
`- which were to be immediately frozen, shipped on dry ice to the Moscow laboratory, and stored
`' at -70 degrees Celcius until the embargo was lifted, at which point the samples were shipped
`on dry ice to the central laboratory. Based On stability testing performed by the sponsor and
`other data published in peer—reviewed literature, the sponsor concluded that the freezing,
`storage, and thawing of these samples would not impact the reliability of the measured HbAlc
`and glucose. Approximately 30% of patients in monotherapy study —038 and 22% of patients
`in the initial combination with metformin trial had at least one frozen HbAlc sample.
`However, using the frozen samples in the calculation (of the primary efficacy endpoint yielded
`virtually identical results to a sensitivity analysis that treated frozen samples as missing.
`Therefore,the frozen samples are included in the efficacy analyses summarized in this
`memorandum.
`
`Demographics: Dr. Lowy and Ms. Mele discuss the patient demographics in detail. Briefly,
`the mean age across the six phase 3 trials was approximately 55 years. Most patients (82-89%)
`were <65 years old. Men and women were generally equally represented, although
`monotherapy study -038 and the add—on to sulfonylurea trials had a slight (55%) female
`predominance. Most patients were Caucasian (55%—8-5%) with blacks comprising 2-7% of the
`randomized patients. Asian representation was reasonable in four of the phase 3 trials
`(monotherapy -03 8, add-on to thiazolidinedione, add-on to sulfonylurea, initial combination
`with metfonnin), ranging from 16-3 5% of randomized patients but Asians accounted for only
`3-5% in the remaining two trials. As expected, mean duration of diagnosed diabetes was
`Shortest in the three trials enrolling treatment-nai'Ve patients (2-3 years) and longest for the
`three add-on combination therapy trials (5-7 years). Mean body ma