`
`
`
`
`
`
`CENTER FOR DRUG EVALUATION AND
`RESEARCH
`
`
`
`APPLICATION NUMBER:
`
`201281Orig1s000
`
`STATISTICAL REVIEW(S)
`
`
`

`

`
`
`
`
`U.S. Department of Health and Human Services
`Food and Drug Administration
`Center for Drug Evaluation and Research
`Office of Translational Sciences
`Office of Biostatistics
`
`
`Statistical Review and Evaluation
`CLINICAL STUDIES
`
`Applicant:
`Date(s):
`Review Priority:
`
`Biometrics Division:
`Statistical Reviewer:
`Concurring Reviewers:
`
`NDA/BLA Serial Number: NDA 201281 / Sequence 0000
`Drug Name:
`Linagliptin/Metformin Hydrochloride Tablets
`Indication(s):
`To improve glycemic control in adults with type 2 diabetes
`mellitus as an adjunct to diet and exercise
`Boehringer Ingelheim Pharmaceuticals, Inc.
`January 19, 2011
`Standard (10-month)
`
`Division of Biometrics 2 (HFD-715)
`Wei Liu, Ph.D.
`J. Todd Sahlroot, Ph.D. (Deputy Director)
`
`Metabolism and Endocrinological Products (HFD-510, DMEP)
`Hyon Kwon, M.D.
`Ilan Irony, M.D. (Team Leader)
`
`
`
`Medical Division:
`Clinical Team:
`
`Project Manager:
`
`
`Keywords: NDA review, clinical studies, factorial design
`
`Reference ID: 3029696
`
`

`

`Table of Contents
`LIST OF TABLES (OPTIONAL)..............................................................................................................................3
`LIST OF FIGURES (OPTIONAL)............................................................................................................................3
`1.
`EXECUTIVE SUMMARY ................................................................................................................4
`TABLE 1.1. GLYCEMIC PARAMETER HBA1C AT WEEK 24 FOR LINAGLIPTIN AND METFORMIN, ALONE AND IN
`COMBINATION IN PATIENTS WITH TYPE 2 DIABETES (LOCF).................................................................4
`INTRODUCTION ..............................................................................................................................6
`OVERVIEW ..............................................................................................................................................6
`DATA SOURCES.......................................................................................................................................7
`STATISTICAL EVALUATION .......................................................................................................8
`DATA AND ANALYSIS QUALITY ..............................................................................................................8
`EVALUATION OF EFFICACY .....................................................................................................................9
`EVALUATION OF SAFETY.......................................................................................................................19
` FINDINGS IN SPECIAL/SUBGROUP POPULATIONS ...........................................................21
`SUMMARY AND CONCLUSIONS ...............................................................................................25
`STATISTICAL ISSUES AND COLLECTIVE EVIDENCE................................................................................25
`5.1
`CONCLUSIONS AND RECOMMENDATIONS..............................................................................................26
`5.2
`LABELLING COMMENTS ........................................................................................................................27
`5.3
`APPENDIX I. TIME COURSES OF HBA1C CHANGES FROM BASELINE BETWEEN ACTIVE
`TREATMENTS AND PLACEBO. .................................................................................................29
`
`2.1
`2.2
`
`3.1.
`3.2.
`3.3.
`
`2.
`
`3.
`
`4.
`5.
`
`APPENDIX II. FOREST PLOTS OF HBA1C CHANGES FROM BASELINE BETWEEN ACTIVE
`TREATMENTS AND PLACEBO IN SUBGROUPS AT WEEK 24...........................................30
`SIGNATURES/DISTRIBUTION LIST (OPTIONAL)..........................................................................................33
`CHECK LIST ...........................................................................................................................................................34
`
`
`
`Reference ID: 3029696
`
`2
`
`

`

`
`LIST OF TABLES (Optional)
`Table 2.1.1: List of Studies Designed to Assess Safety and Efficacy .......................................................................7
`Table 3.2.1. Patient disposition and demographic information.............................................................................11
`Table 3.2.2. Glycemic Parameters (HbA1c) at Week 24 for Linagliptin and Metformin, Alone and in
`Combination in Patients with Type 2 Diabetes..................................................................................15
`Table 3.2.3. Glycemic Parameters (Fasting Plasma Glucose) at Week 24 for Linagliptin and Metformin,
`Alone and in Combination in Patients with Type 2 Diabetes ...........................................................16
`Table 3.3.1. List of Adverse Events by Treatments on All Randomized Patients with Type 2 Diabetes ...........19
`Table 3.3.2. List of Laboratory Assays That Were Significantly Worse on Patients Treated by Lina 2.5 mg
`+ Met 500 mg Twice Daily Versus At Least One Component ..........................................................20
`Table 3.3.3. List of Laboratory Assays That Were Significantly Worse on Patients Treated by Lina 2.5 mg
`+ Met 1000 mg Twice Daily Versus At Least One Component........................................................20
`
`
`
`
`
`
`LIST OF FIGURES (Optional)
`Figure 3.2.1. Overview of the study design..............................................................................................................10
`Figure 3.2.2. Baseline Levels of HbA1c in Different Treatment Groups. .............................................................12
`Figure 3.2.3. HbA1c Changes from Baseline Over 24 Weeks with Linagliptin + Metformin, Alone and in
`Combination in Study 46. .................................................................................................................17
`Figure 3.2.4. The Plot of HbA1c Changes from Baseline versus Baseline Levels A: Linagliptin 5 mg and
`placebo, B: Metformin 500mg and placebo, C: Linagliptin 2.5 mg+Metformin 500mg and
`placebo, D: Metformin 1000mg and placebo, and E: Linagliptin 2.5 mg+Metformin
`1000mg and placebo, respectively, in Study 46 to Week 24 (LOCF). ............................................18
`Figure 4.1.2. The Plot of HbA1c Changes from Baseline versus Baseline Levels between Linagliptin 2.5
`mg+Metformin 1000mg Twice Daily and Metformin 1000mg Twice Daily Treatments in
`Study 46 at Week 24...........................................................................................................................22
`Figure 4.1.3. The Plot of HbA1c Changes from Baseline versus Baseline Levels between Linagliptin 2.5
`mg+Metformin 500mg Twice Daily and Linagliptin 5 mg Once Daily Treatments in Study
`46 at Week 24......................................................................................................................................23
`Figure 4.1.4. The Plot of HbA1c Changes from Baseline versus Baseline Levels between Linagliptin 2.5
`mg+Metformin 500mg Twice Daily and Metformin 500mg Twice Daily Treatments in
`Study 46 at Week 24...........................................................................................................................24
`
`
`
`
`
`Reference ID: 3029696
`
`3
`
`

`

`EXECUTIVE SUMMARY
`
`
`1.
`
`This statistical review covers one randomized trial of co-administered linagliptin and metformin
`(Study 46). Other (Lina + Met) combination trials submitted by the sponsor (Studies 17, 18 and
`20) were reviewed in NDA 202180, the original submission for linagliptin, therefore were not
`reviewed as part of the current submission.
`
`Confirmation of efficacy: The results of the pivotal study 1218.46 support the efficacy of
`linagliptin add-on to metformin hydrochloride at fixed dose as an adjunct to diet and exercise to
`improve glycemic control in adults with type 2 diabetes mellitus after 24 weeks of treatment
`based HbA1c reduction. Particularly, the combination treatment is statistically superior to the
`placebo and to each corresponding component treatment after 24 weeks treatment at a 0.05 level
`(two-sided).
`
`
`Table 1.1. Glycemic Parameter HbA1c at Week 24 for Linagliptin and Metformin,
`Alone and in Combination in Patients with Type 2 Diabetes (LOCF)
`
`
`
`Study population
`
`Placebo
`
`Lina 5 mg
`Once Daily
`
`Met 500 mg
`Twice Daily
`
`n = 65
`8.7 (0 1)
`0.1 (0.1)
`--
`
`
`7 (10.8)
`
`n = 135
`8.7 (0.1)
`-0.5 (0.1)
`-0.6 (-0.8, -0.3)
`
`
`14 (10.4)
`
`n = 141
`8.7 (0.1)
`-0.6 (0.1)
`-0.8 (-1.1, -0.5)
`
`
`27 (19.1)
`
`Lina 2.5 mg +
`Met 500 mg
`Twice Daily
`n = 137
`8.7 (0.1)
`-1.2 (0.1)
`-1.3 (-1.6, -1.1)
`-0.6 (-0.8, -0.4)
`-0.8 (-1.0, -0.6)
`44 (32.1)
`
`Met 1000 mg
`Twice Daily
`
`n = 138
`8.5 (0.1)
`-1.0 (0.1)
`-1.2 (-1.5, -0.9)
`
`
`43 (31.6)
`
`Lina 2.5 mg +
`Met 1000 mg
`Twice Daily
`n = 140
`8.7 (0.1)
`-1.6 (0.1)
`-1.7 (-2.0, -1.4)
`-0.5 (-0.7, -0.3)
`-1.1 (-1.4, -0.9)
`76 (54.3)
`
`Number of patients
`Baseline (mean, SE)
`Change from baseline1 (SE)
`Diff from placebo1 (95% CI)
`Diff from Met alone1 (95% CI)
`Diff from Lina alone1 (95% CI)
`achieving A1C <7% (n, %)*
`Patients (%, n) receiving rescue
`29.2 (19)
`11 1 (15)
`medication
`(* the numbers were based on LOCF population)
`
`The results from the sensitivity analyses (such as MMRM, completers, and per protocol) and key
`secondary endpoint, fasting plasma glucose level, support the superior of the combination to the
`placebo and to each corresponding component treatment on both HbA1c and FPG reductions
`after 24 weeks treatment at a 0.05 level (two-sided).
`
`Subgroup analyses suggest that females derive greater benefit from adding either Lina or Met to
`the other drug than do males.
`
`There were no significant differences in adverse event rates between each (Lina+Met)
`combination and its components. Laboratory assays suggest significant elevations in some
`immune system reactions in patients treated by the combined (Lina+Met) drugs versus those by
`the component drugs.
`
`The results from non-LOCF analysis methods (this reviewer’s MMRM, completers, and per
`protocol) showed that linagliptin 5 mg was not statistically superior to placebo at the 0.05 alpha
`
`
`13.5 (19)
`
`7.3 (10)
`
`8.0 (11)
`
`4.3 (6)
`
`Reference ID: 3029696
`
`4
`
`

`

`
`
`level (two-sided). Estimated treatment differences for HbA1c were -0.1 or -0.2. These smaller
`treatment differences compared to the LOCF results were due primarily to differences in the
`estimated changes from baseline in the placebo group which were much lower (i.e., greater
`improvement) in the sensitivity analyses. It is noted also in the LOCF population that the
`percentages of 7% HbA1c responders (see Table 1 above) for placebo (10.8%) and linagliptin
`(10.4%) were identical raising questions about the efficacy of linagliptin 5 mg QD in study 46.
`
`This finding is, however, not critical to the determination of efficacy of the combinations
`(Lina+Met) since the determination does not require efficacy data from placebo and the data
`from linagliptin monotherapy is used only to support the efficacy of metformin in the
`combination. Nevertheless the efficacy of linaglipin monotherapy needs to be considered in the
`context of the submission.
`
`
`Labelling Recommendations:
`
`The statistical review addresses statements in the label (section 14.1) concerning:
`1. Description of randomization: The sponsor should state that “Randomization was stratified
`by baseline HbA1c (<8.5% versus (cid:149)8.5%) and number of prior oral anti-diabetic drug (none
`versus monotherapy).”
`
`
`2.
`
`
`3.
`
` In the third paragraph of section 14.1, the sponsor should indicate that at these results were
`based the analyses using the last observation carried forward (LOCF) method.
`
`Subgroup of patients with high baseline (14.1 paragraph 4): this claim was not supported by
`data. The mean reduction from baseline in A1c were also greater for patients with higher
`baseline A1c in the placebo group (see review Figure 3.2.4 A-D). The differences between
`strata were not significant; and the trends of differences between patients stratified using
`baseline A1c cutoff 8.5 in subgroup analyses (review Figures 4.1.1-4.1.4) varied.
`
`
`4. Efficacy results of open label arm (14.1, line 648-651): these results are not valid for
`efficacy claim because of no placebo or active comparator group. As seen in the comment
`#3 above, the mean reduction from baseline in HbA1c were also greater for patients with
`higher baseline A1c in the placebo group (see review Figure 3.2.4 A-D).
`
`
`5.
`
`6. Efficacy data for extension: HbA1c is not a primary endpoint of this study. The interim
`analysis results listed in the label were not representative because they were based only on a
`very small portion of the patients enrolled in the study: 10 (6%) in lina 2.5 mg/met 500 mg
`twice daily group; 10 (4%) in lina 2.5 mg/met 1000 mg twice daily group; and 9 (5%) in
`met 1000 mg twice daily group. These efficacy data (HbA1c and FPG) should not be
`included in the label prior to the extension study completion.
`
`Figure 1 should be a plot of completers.
`
`
`7.
`
`In Table 7, the upper 95% CI of “Difference from placebo” for Metformin 1000 mg twice
`daily should be “-0.9”.
`
`
`
`Reference ID: 3029696
`
`5
`
`

`

`
`2.
`
`
`
`INTRODUCTION
`
`2.1 Overview
`
`
`Linagliptin is an inhibitor of DPP-4, an enzyme which rapidly degrades incretin hormones
`glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP),
`thereby increasing insulin release and decreasing the level of glucagon in the circulation in a
`glucose-dependent manner. Metformin was approved for patients with type 2 diabetes in March
`1995, subject of NDA 020357. The efficacy and safety of linagliptin as a monotherapy or as an
`add-on treatment to other diabetic drugs were reviewed in NDA 201280. The current submission
`is specifically focused on linagliptin add-on to metformin hydrochloride at fixed dose as an
`adjunct to diet and exercise to improve glycemic control in adults with type 2 diabetes mellitus.
`To support the proposed draft labeling for linagliptin/metformin hydrochloride fixed dose
`combination tablets, this NDA references the approved US labeling for Glucophage (metformin
`hydrochloride) tablets.
`
`The applicant submitted data from one phase 2 study 1218.06, four phase 3 studies 1218.17,
`1218.18, 1218.20, and the pivotal study 1218.46, and two open-label extension studies 1218.40
`and 1218.52. In total, 3084 patients with type 2 diabetes received treatment with linagliptin and
`metformin; of these, 2585 were treated for at least 24 weeks and 1749 for at least 52 weeks. For
`the assessment of efficacy and safety, additional supportive non-clinical and clinical studies are
`cross- referenced to Modules 4 and 5 of NDA 201280 for linagliptin tablets. The phase 3 studies
`whose results are shown in the sponsor proposed label are marked in the shaded area of Table
`2.1.1. Since the data for studies 1218.17, 1218.18, and 1218.20 were reviewed in NDA 201280,
`the data of the pivotal study 1218.46 (open-label arm excluded) is selected for full statistical
`review and evaluation in this review.
`
`
`
`
`Reference ID: 3029696
`
`6
`
`

`

`Table 2.1.1: List of Studies Designed to Assess Safety and Efficacy
`Study
`Phase and Design
`Treatment
` # of Subjects per Arm
`Study Population
`Period
`12 weeks
`
`Lina 1 mg + Met (65)
`Lina 5 mg + Met (66)
`Lina 10 mg + Met (66)
`PBO + Met (71)
`
`Met failure
`
`
`24 weeks
`
`24 weeks
`
`
`Lina 5 mg + Met (523)
`PBO + Met (177)
`
`
`Met failure
`
`Lina 5 mg + Met + SU (792)
`PBO + Met + SU (263)
`
`Met + SU failures
`
`1218.06
`
`Phase 2
`Double-bind placebo-
`controlled, Met add-on
`(also include an open-
`label Glm arm)
`
`
`1218.17 Phase 3
`Double-bind placebo-
`controlled
`1218.18 Phase 3
`Bouble-bind placebo-
`controlled
`1218.20 Phase 3
`Double-bind active-
`controlled
`1218.40 Phase 3
`Open-label long-term
`extension
`
`52 weeks*
`
`Lina 5 mg + Met (778)
`Glm + Met (781)
`
`78 weeks*
`
`Lina 5 mg (2121)
`
`Met failures
`
`in patients with Type 2
`diabetics who
`continued their
`treatment from studies
`1218.15, 1218.16,
`1218.17, and 1218.18
`T2DM patients with
`insufficient glycaemic
`control either drug-
`naïve or treated with
`one oral antidiabetic
`agent
`
`Patients who had
`completed trial 1218.46
`and were not being
`treated with rescue
`medication
`
`1218.46 Phase 3
`Pivotal double-bind
`placebo-controlled
`
`24 weeks
`
`1218.52 Phase 3
`Double-bind parallel
`group extension
`
`PBO (72)
`Lina 2.5 mg+Met 500 mg, twice
`daily (143)
`Lina 2.5 mg+Met 1000 mg, twice
`daily (143)
`Lina 2.5 mg+Met 1000 mg, twice
`daily (66)^
`Lina 5 mg once daily (142)
`Met 500 mg twice daily (144)
`Met 1000 mg twice daily (147)
`54 weeks Lina 2.5 mg + Met 500 mg, twice
`daily (225)
`Lina 2.5 mg + Met 1000 mg,
`twice daily (171)
`Met 1000 mg, twice daily (170)
`
`
`Lina = linagliptin, PBO = placebo, Pio = pioglitazone, Met = metformin ((cid:149)1500 mg/day or maximum tolerated
`dose), SU = sulfonylurea (maximum tolerated), Glm= Glimepiride (1 to 4 mg/day)
`* 104 weeks. On going. Data are available in interim analysis at 52 weeks
`^ open-label
`
`2.2 Data Sources
`
`
`
`
`
`Reference ID: 3029696
`
`7
`
`

`

`The sponsor submitted this NDA including the study data to the FDA CDER Electronic
`Document Room (EDR). The submission is recorded in the EDR with the link shown below. The
`data were submitted in SAS Xport transport format.
`
`
`Application:
`
`NDA201281/0000
`
`Company
`
`Boehringer Ingelheim
`
`Drug
`
`Linagliptin
`
`CDER EDR link
`
`\\CDSESUB1\EVSPROD\ NDA201281\0000
`
`Letter date
`
`1/19/2011
`
`
`The applicant’s electronic submission was well-organized. Parallel structure in the presentation
`of the results across all studies was well-done and appreciated by the reviewer.
`
`All graphs and tables in the review were created by this reviewer unless otherwise noted.
`
`3.
`
`
`STATISTICAL EVALUATION
`
`3.1.
`
`Data and Analysis Quality
`
`I reviewed the quality and integrity of the submitted data. Relevant issues include:
`• Whether it is possible to reproduce the primary analysis dataset from tabulation or “raw”
`datasets : yes
`• Whether it is possible to trace how the primary endpoint was derived from the original
`data source (e.g., case report form): yes.
`• Whether it is possible to verify the randomized treatment assignments: yes
`• Findings from the Division of Scientific Investigation or other source(s) that question the
`usability of the data:
`Susan Leibenhaut, MD, from the Division of Scientific Investigations requested to verify
`the following information:
`Study 1218.46: the number of treated subjects at the following India sites:
`•at site 91004 there are 24 treated subjects
`•at site 91015 there are 30 treated subjects.
`
`This reviewer checked the ADSL.xpt and got the following results:
`Country=INDIA
`Site 91004: 24 (randomized), 28 (safety population), 21 (FAS population), 2 rescued
`(with sulphonylurea)
`
`
`
`Reference ID: 3029696
`
`8
`
`

`

`Site 91015: 30 (randomized), 33 (safety population), 30 (FAS population), 8 rescued (6
`with sulphonylurea and 2 with metformin)
`
`
` Obs USUBJID RESCUE
`
` 243 1218-0046-047321 SULPHONYLUREA
` 713 1218-0046-049642 SULPHONYLUREA
` 723 1218-0046-049658 SULPHONYLUREA
` 724 1218-0046-049659 SULPHONYLUREA
` 762 1218-0046-049776 METFORMIN
` 764 1218-0046-049778 SULPHONYLUREA
` 766 1218-0046-049780 SULPHONYLUREA
` 832 1218-0046-049901 METFORMIN
`
`Therefore the numbers of randomized subjects agree with the number of treated subjects listed in
`the table.
`
`I did not encounter any problem or difficulty to process the data.
`
`
`
`3.2.
`
`Evaluation of Efficacy
`
`
`
`Study Design and Endpoints
`
`Study 46 is a Phase III randomised, double-blind, placebo-controlled parallel group study to
`compare the efficacy and safety of twice daily administration of the free combination of
`linagliptin 2.5 mg + metformin 500 mg or of linagliptin 2.5 mg + metformin 1000 mg, with the
`individual components of metformin (500 mg or 1000 mg, twice daily) and linagliptin (5 mg,
`once daily) over 24 weeks in drug naïve or previously treated (4 weeks washout and 2 weeks
`placebo run-in) type 2 diabetic patients with insufficient glycaemic control.
`
`Study 46 was a multi-national, multi-centre trial with 133 sites in 14 countries (Canada, Croatia,
`Estonia, France, Germany, India, Lithuania, Mexico, Romania, Russia, Sweden, The
`Netherlands, Tunisia, and Ukraine)
`
` A
`
` total of 1770 patients were enrolled into this study and 792 patients were randomized in a
`1:2:2:2:2:2 ratio to either placebo (72 patients), linagliptin 5 mg (142 patients), metformin 500
`mg (144 patients), metformin 1000 mg (147 patients), linagliptin 2.5 mg + metformin 500 mg
`(143 patients), or linagliptin 2.5 mg + metformin 1000 mg (143 patients). The sample sizes were
`determined using two-sided t-tests at alpha=0.05, standard deviation and effect size for HbA1c
`change from baseline to 24 weeks equal to 1.1% and -0.5% (against placebo), and power 0.85.
`Randomization was stratified by baseline HbA1c (<8.5% versus (cid:149)8.5%) and number of prior oral
`anti-diabetic drug (PAD, none versus monotherapy). The main reason for non-randomization
`was in-/exclusion criteria not met (42.8% of the enrolled patients). All of the randomized
`patients were treated. The most frequent reasons for discontinuation were due to adverse events
`(3.3%), refusal to continue trial medication (2.9%), and lack of efficacy (2.3%).
`
`The sponsor’s design diagram of the study 1218.46 is shown in Figure 3.2.1.
`
`
`
`9
`
`Reference ID: 3029696
`
`

`

`
`Figure 3.2.1. Overview of the study design.
`
`
`
`
`
`1 Patients who received 1000 mg metformin had to undergo a 2-week forced titration
`
`The primary endpoint for study 1218.46 is the HbA1c change from baseline to 24 weeks. The
`HbA1c levels were measured at weeks 0, 6, 12, 18, and 24.
`
`The key secondary endpoints (and other endpoints) include the change from baseline in fasting
`plasma glucose (FPG), occurrence of treat-to-target response (i.e. HbA1c on treatment <7.0% or
`<6.5%), occurrence of relative efficacy response (i.e. HbA1c lowering by 0.5%), change from
`baseline in two-hour postprandial glucose (2hPPG) for Meal Tolerance Test (MTT), and the use
`of rescue medication.
`
`
`Patient Disposition, Demographic and Baseline Characteristics
`
`A description of the patient populations in the review is shown in Table 3.2.1.
`
`
`
`
`
`
`10
`
`Reference ID: 3029696
`
`

`

`
`Table 3.2.1. Patient disposition and demographic information
`
`Study population
`Placebo
`Lina 5 mg
`Met 500 mg
`Lina 2.5 mg
`Met
`Once Daily*
`Twice Daily
`Twice Daily* +
`1000 mg
`Met 500 mg
`Twice Daily
`Twice Daily
`143
`137 (96%)
`133 (93%)
`125 (87%)
`10 (7%)
`
`
`56 (1)
`30-80
`22%
`
`51%
`
`72%
`
`Randomized
`
`FAS
`
`Per Protocol
`
`Completers
`
`Rescued
`
`72
`65 (90%)
`63 (87%)
`52 (72%)
`19 (26%)
`
`
`56 (1)
`33-78
`21%
`
`50%
`
`64%
`
`142
`135 (95%)
`130 (92%)
`118 (83%)
`15 (11%)
`
`
`56 (1)
`28-77
`23%
`
`56%
`
`68%
`
`144
`141 (98%)
`140 (97%)
`120 (83%)
`19 (13%)
`
`
`53 (1)
`30-73
`16%
`
`57%
`
`65%
`
` 4
`
` (3%)
`49 (33%)
`70 (48%)
`7 (5%)
`17 (12%)
`75 (51%)
`
` 7
`
` (5%)
`37 (26%)
`73 (51%)
`9 (6%)
`17 (12%)
`76 (53%)
`
` 7
`
` (5%)
`49 (34%)
`64 (44%)
`7 (5%)
`17 (12%)
`75 (52%)
`
` 5
`
` (4%)
`43 (30%)
`73 (51%)
`7 (5%)
`14 (10%)
`79 (56%)
`
` 4
`
`
`
`Reference ID: 3029696
`
`11
`
`Lina 2.5 mg
`Twice Daily* +
`Met 1000 mg
`Twice Daily
`143
`140 (98%)
`135 (94%)
`129 (90%)
`6 (4%)
`
`
`56 (1)
`26-77
`24%
`
`54%
`
`66%
`
` 5
`
` (4%)
`45 (31%)
`65 (45%)
`10 (7%)
`18 (13%)
`77 (54%)
`
`147
`138 (94%)
`129 (88%)
`122 (83%)
`12 (8%)
`
`
`55 (1)
`25-76
`21%
`
`53%
`
`65%
`
`
`Age (yr)
`Mean(SE)
`Range
`% (cid:149)65 yr
`Gender
`% males
`Race
`% White
`Region
`Africa
` (6%)
`Asia
`26 (36%)
`Europe
`30 (42%)
`2 (3%)
`N. Am*
`10 (14%)
`S. Am^
`PAD
`38 (53%)
`*Total daily dose of linagliptin is equal to 5 mg
`* Canada
`^ Mexico
`PAD: Previous antidiabetic medication used
`
`The baseline levels of HbA1c in the five double-bind treatment groups are compared in boxplots
`as shown in Figure 3.2.2. In each boxplot the bottom and top of the box are the 25th and 75th
`percentile, respectively; the “+” and the line near the middle of the box is the mean and median
`(50th percentile), respectively; the top line above the box is the maximum observation; and the
`bottom line below the box is the minimum observation. Across the different treatment groups,
`the baseline levels of HbA1c appear to have similar means and comparable variability.
`
`

`

`Figure 3.2.2. Baseline Levels of HbAlc in Different Treatment Groups.
`
`14
`
`10
`
`12
`
`HBAC
`
`AA005
`
`AE007
`
`AEOOB
`
`AM501
`
`AM502
`
`P
`
`TRTP
`
`AA005: Linagliptin 5 mg once daily
`AE007: Linagliptin 2.5 mg twice daily + Metformin 500 mg twice daily
`AE008: Linagliptin 2.5 mg twice daily + Metfomiin 1000 mg twice daily
`AM501: Metfonnin 500 mg twice daily
`AM502: Metformin 1000 mg twice daily
`P: Placebo
`
`Statistical Methodologies
`
`The sponsor’s primary analysis is to test the superiority of the combination therapies over the
`monotherapies in a hierarchical manner with an analysis of covariance (ANCOVA) method with
`the treatment and previous anti-diabetes therapy as fixed classification effects and baseline
`HbAlc as linear covariate. The patient population for the primary analysis is the full analysis set
`(FAS) that was a subset of the treated set including all patients who had a baseline and at least
`one on-treatment HbAlc measurement available. The last observation carried forward (LOCF)
`approach was used to replace missing data. This analysis method is adequate for demonstrating
`efficacy for a combination product.
`
`12
`
`Reference ID: 3029696
`
`

`

`The sponsor’s secondary analysis used ANCOVA (with LOCF) for the continuous variables, and
`descriptive statistics and logistic regression for the binary response variables. For safety
`endpoints, the descriptive statistics and Kaplan-Meier analysis were used.
`
`The sponsor imputed data in the following cases: (1) if a patient received rescue medication
`before measurement of the first on-treatment HbA1c value, the baseline HbA1c value was
`carried forward; and (2) missing values within a course of measurements on treatment were
`interpolated based on the last observed value before the missing visit and the first observed value
`after the missing visit. This reviewer evaluated the datasets and found that in case (1) the number
`of BOCF was small, therefore the effect of above method on the primary analysis using LOCF
`can be ignored.
`
`This reviewer’s statistical analysis methods have changed slightly from that used in NDA
`201280 as specified below. I used LOCF on the sponsor’s ANCOVA model as the main
`imputation method in the primary analysis and in subgroup analysis. The methods for sensitivity
`analyses were the same as in the review of NDA 201280, that is, I used the per protocol and
`completers populations for sensitivity analysis. I used MMRM as a secondary analysis with an
`additional fixed effect ‘visit week’ to the general model applied to the original dataset. The
`completers were used for longitudinal graphs.
`
`
`
`Results and Conclusions
`
`The superiority of linagliptin and metformin, alone and in combination over placebo was tested
`for HbA1c change from baseline to week 24 at the level of (cid:302)=0.05 (two-sided) on different
`analysis populations. The treatment differences between an anti-diabetic drug and placebo,
`calculated as the adjusted mean change in HbA1c from baseline at Week 24, are summarized in
`Table 3.2.2 for the primary and sensitivity analyses.
`
`Testing each treatment of anti-diabetic drug(s) over the placebo using FAS with LOCF, the
`applicant's results suggested significant reduction in HbA1c from the baseline level after 24
`weeks treatment. These results were verified by this reviewer using the sponsor’s model and
`method. The (Lina + Met) combinations have larger reductions from placebo in HbA1c levels
`than that by either component alone, suggesting additional effects from each component drug.
`Additional efficacy was also seen using three other analysis methods by this reviewer, namely
`the MMRM method and the two ANCOVA sensitivity analyses using per protocol population
`and completers populations. However, unlike the results obtained using the FAS population with
`LOCF, the three analyses by this reviewer revealed that lina 5 mg was not superior over the
`placebo at the 0.05 level (two-sided).
`
`The results for the secondary endpoint fasting plasma glucose (FPG) compared with placebo are
`listed in Table 3.2.3. As seen in HbA1c, the additional efficacy of the (Lina + Met) combinations
`as compared to either component was also observed in the analysis results of FPG using FAS
`population with LOCF. Again, the three analyses (the MMRM method and the two ANCOVA
`sensitivity analyses using per protocol population and completers) by this reviewer also revealed
`
`
`
`Reference ID: 3029696
`
`13
`
`

`

`that lina 5 mg was not superior over the placebo at the 0.05 level (two-sided) for the secondary
`endpoint FPG, different from the results obtained using the FAS population with LOCF.
`
`In summary, the (Lina + Met) combination treatment arms were statistically superior to the
`placebo and to each corresponding component treatment on both HbA1c and FPG reductions
`after 24 weeks treatment at a 0.05 level (two-sided).
`
`
`
`Reference ID: 3029696
`
`14
`
`

`

`Table 3.2.2. Glycemic Parameters (HbA1c) at Week 24 for Linagliptin and
`Metformin, Alone and in Combination in Patients with Type 2 Diabetes
`
`
`Study population
`
`Placebo
`
`Lina 5 mg
`Once Daily*
`
`Met 500 mg
`Twice Daily
`
`Lina 2.5 mg +
`Met 500 mg
`Twice Daily
`
`Met 1000 mg
`Twice Daily
`
`Lina 2.5 mg +
`Met 1000 mg
`Twice Daily
`
`n = 137
`8.7
`-1.2
`-1.3 (-1.6, -1.1)
`41 (30.1)
`
`n = 138
`8.5
`-1.1
`-1.2 (-1.5, -0.9)
`42 (30.7)
`
`n = 140
`8.7
`-1.6
`-1.7 (-2.0, -1.4)
`74 (53.6)
`
`7.3
`
`8.0
`
`4.3
`
`n = 137
`8.7 (0.1)
`-1.2 (0.1)
`-1.3 (-1.6, -1.1)
`-0.6 (-0.8, -0.4)
`-0.8 (-1.0, -0.6)
`44 (32.1)
`
`n = 138
`8.5 (0.1)
`-1.0 (0.1)
`-1.2 (-1.5, -0.9)
`
`
`43 (31.6)
`
`n = 140
`8.7 (0.1)
`-1.6 (0.1)
`-1.7 (-2.0, -1.4)
`-0.5 (-0.7, -0.3)
`-1.1 (-1.4, -0.9)
`76 (54.3)
`
`7.3 (10)
`
`8.0 (11)
`
`4.3 (6)
`
`n = 137
`8.7 (0.1)
`-1.3 (0.1)
`-1.0 (-1.3, -0.6)
`-0.5 (-0.7, -0.3)
`-0.7 (-0.9, -0.5)
`
`n = 135
`8.5 (0.1)
`-1.1 (0.1)
`-0.8 (-1.1, -0.6)
`
`
`
`n = 139
`8.6 (0.1)
`-1.7 (0.1)
`-1.4 (-1.7, -1.1)
`-0.5 (-0.8, -0.3)
`-1.2 (-1.4, -0.9)
`
`125
`8.7 (0.1)
`-1.3 (0.1)
`-0.9 (-1.1, -0.6)
`-0.5 (-0.7, -0.2)
`-0.7 (-0.9, -0.5)
`
`122
`8.5 (0.1)
`-1.2 (0.1)
`-0.8 (-1.1, -0.6)
`
`
`
`129
`8.6 (0.1)
`-1.7 (0.1)
`-1.3 (-1.6, -1.1)
`-0.5 (-0.7, -0.3)
`-1.2 (-1.4, -1.0)
`
`122
`8.7 (0.1)
`-1.3 (0.1)
`-0.9 (-1.2, -0.6)
`-0.4 (-0.6, -0.2)
`-0.7 (-0.9, -0.5)
`
`118
`8.5 (0.1)
`-1.2 (0.1)
`-0.9 (-1.2, -0.6)
`
`
`
`125
`8.7 (0.1)
`-1.7 (0.1)
`-1.3 (-1.6, -1.0)
`-0.4 (-0.7, -0.2)
`-1.2 (-1.4, -1.0)
`
`15
`
`n = 141
`8.7
`-0.6
`-0.8 (-1.0, -0.5)
`26 (18.6)
`
`29.2
`
`11.1
`
`13.5
`
`n = 141
`8.7 (0.1)
`-0.6 (0.1)
`-0.8 (-1.1, -0.5)
`
`
`27 (19.1)
`
`29.2 (19)
`
`11 1 (15)
`
`13.5 (19)
`
`Full Analysis Set (LOCF): reported by applicant
`Number of patients
`n = 65
`n = 135
`Baseline (mean)
`8.7
`8.7
`Change from baseline1
`0.1
`-0.5
`Diff from placebo1 (95% CI)
`--
`-0.6 (-0.9, -0.3)
`achieving A1C <7% (n, %)
`7 (10.8)
`14 (10.4)
`Patients (%) receiving rescue
`medication
`Full Analysis Set (LOCF): this reviewer’s analysis
`Number of patients
`n = 65
`n = 135
`Baseline (mean, SE)
`8.7 (0 1)
`8.7 (0.1)
`Change from baseline1 (SE)
`0.1 (0.1)
`-0.5 (0.1)
`Diff from placebo1 (95% CI)
`--
`-0.6 (-0.8, -0.3)
`Diff from Met alone1 (95% CI)
`
`
`Diff from Lina alone1 (95% CI)
`
`
`achieving A1C <7% (n, %)
`7 (10.8)
`14 (10.4)
`Patients (%, n) receiving rescue
`medication
`Full Analysis Set: this reviewer’s analysis (MMRM, Original data)
`Number of patients
`n = 64
`n = 135
`n = 136
`Baseline (mean, SE)
`8.7 (0 1)
`8.7 (0.1)
`8.6 (0.1)
`Change from baseline1 (SE)
`-0.3 (0.1)
`-0.5 (0.1)
`-0.8 (0.1)
`Diff from placebo1 (95% CI)
`--
`-0.2 (-0.5, 0.1)
`-0.4 (-0.7, -0.1)
`Diff from Met alone1 (95% CI)
`
`
`
`Diff from Li