FINAL BRIEFING DOCUMENT FOR ADVISORY COMMITTEE MEETING

Efficacy Review Of New Drug Application

 

 (Serial Number)                            21487 (000)

Sponsor:                                              Forest Laboratories

Drug:                                                   Memantine

Proposed Indication:                           Alzheimer’s Disease 

Material Submitted:                             New Drug Application

Correspondence Date:                         12/19/02

Date Received / Agency:                      12/24/02

Date Review Completed                       8/19/03

Reviewer:                                             Ranjit B. Mani, M.D.

1        Table Of Contents

1      Table Of Contents                                                                            1

2      Executive Summary                                                                           6

2.1        Recommendation                                                                                                                                  6

2.2        Proposed Indication                                                                                                                             6

2.3        Summary Of Clinical Findings                                                                                                          6

2.3.1     MRZ 9605                                                                                                                                           6

2.3.2     MEM-MD-02                                                                                                                                     8

2.3.3     MRZ 9403                                                                                                                                           9

2.4        Conclusions                                                                                                                                          12

3      Introduction                                                                                    13

4      Organization Of Submission                                                         13

5      Outline Of Review                                                                            14

6      Chemistry                                                                                          14

7      Proposed Mechanism Of Action                                                   15

8      Summary Of Memantine Pharmacokinetics And Clinical Pharmacology                                                                                        15

9      Rating Scales/Outcome Measures Used In Key Efficacy Studies       16

9.1        Primary Efficacy Variables                                                                                                              16

9.1.1     Severe Impairment Battery                                                                                                              16

9.1.2     Alzheimer's Disease Cooperative Study-Activities of Daily Living (ADCS-ADL): Modified 17

9.1.3     Clinician Interview Based Impression of Change-Plus (CIBIC-Plus)                                         17

9.1.4     Clinician Global Impression of Change (CGI-C)                                                                          17

9.1.5     Behavioral Rating Scale In Geriatric Patients (BGP)                                                                   18

9.2        Secondary Efficacy Variables                                                                                                           19

9.2.1     Mini-Mental Status Examination (MMSE)                                                                                    19

9.2.2     Functional Assessment Staging                                                                                                       19

9.2.3     Neuropsychiatry Inventory                                                                                                              19

9.2.4     Resource Utilization In Dementia                                                                                                  19

9.2.5     G2 Scale                                                                                                                                            19

9.2.6     Instrumental Activities of Daily Living Performance Test  (IADLPT)                                       20

9.2.7     Clinical Global Impression of Severity (CGI-S)                                                                           20

9.2.8     Clinician Global Impression of Change (CGI-C) Benefit/Risk Index                                         20

9.3        Rating Scales Not Used As Efficacy Variables                                                                              20

9.3.1     Hamilton Depression Scale (HDS)                                                                                                20

9.3.2     Global Deterioration Scale (GDS)                                                                                                 21

9.3.3     Hachinski Ischemic Scale (Rosen Modification)                                                                         21

10        Tabular Summary Of Key Efficacy Studies                           21

10.1      Study MRZ 9605                                                                                                                                  21

10.2      Study MRZ  9403                                                                                                                                 22

10.3      Study MEM-MD-02                                                                                                                            23

10.3.1       Least Square Mean Change From Baseline In Severe Impairment Battery                            23

10.3.2       Least Square Mean Change From Baseline In ADCS-ADL                                                     23

11        Study MRZ 9605                                                                               24

11.1      Study Protocol                                                                                                                                     24

11.1.1       Objective                                                                                                                                      24

11.1.2       Design                                                                                                                                          24

11.1.3       Sample Size                                                                                                                                  24

11.1.4       Key Inclusion Criteria                                                                                                                24

11.1.5       Key Exclusion Criteria                                                                                                               24

11.1.6       Concomitant Medications                                                                                                          25

11.1.7       Dosage                                                                                                                                          25

11.1.8       Duration                                                                                                                                       26

11.1.9       Schedule                                                                                                                                       27

11.1.10     Outcome Measures                                                                                                                     28

11.2      Analysis Plan                                                                                                                                       28

11.2.1       General Considerations                                                                                                              28

11.2.2       Study Populations                                                                                                                        28

11.2.3       Demographic And Baseline Characteristics                                                                             29

11.2.4       Drug Compliance                                                                                                                        29

11.2.5       Primary Efficacy Parameters                                                                                                     29

11.2.6       Null And Alternative Hypotheses                                                                                              29

11.2.7       Additional Analyses On Primary Efficacy Parameters                                                            29

11.2.8       Pooling Of Centers                                                                                                                     29

11.2.9       Secondary Efficacy Parameters                                                                                                 30

11.2.10     Responder Analyses                                                                                                                    30

11.2.11     Subgroup Analyses                                                                                                                      30

11.2.12     Handling Of Missing Items                                                                                                        30

11.2.13     Sample Size Rationale                                                                                                                31

11.3      Protocol Amendments                                                                                                                       31

11.4      Actual Analyses Performed                                                                                                              31

11.4.1       Alternative Imputation Schemes For Analysis Of CIBIC-Plus                                               31

11.4.2       Additional Analyses Of The Severe Impairment Battery                                                         31

11.4.3       Elimination Of The Resource Utilization In Dementia Analyses From The Main Study Report                    32

11.4.4       Elimination Of The Treated-Per-Protocol Analyses                                                               32

11.4.5       Elimination Of Subgroup Efficacy Analyses Based On Plasma Levels                                 32

11.4.6       Determination Of The Primary Reason For Discontinuation                                                 32

11.5      Efficacy Results                                                                                                                                   32

11.5.1       Patient Disposition                                                                                                                     32

11.5.2       Treatment Duration                                                                                                                     32

11.5.3       Demographic And Other Baseline  Characteristics                                                                 33

11.5.4       Primary Efficacy Analysis                                                                                                          34

11.5.5       Analysis Of Secondary Efficacy Measures                                                                               37

11.5.6       Additional Sponsor Analyses                                                                                                     38

11.5.7       Agency Subgroup Analysis                                                                                                         39

11.6      Sponsor’s Conclusions Regarding Efficacy                                                                                  40

11.7      Agency Statistical Reviewer’s Comments                                                                                    40

11.8      Reviewer’s Comments                                                                                                                       40

12        Study MRZ 9403                                                                               41

12.1      Title                                                                                                                                                       41

12.2      Objective                                                                                                                                               42

12.3      Design                                                                                                                                                    42

12.4      Duration                                                                                                                                                42

12.5      Dosage                                                                                                                                                   42

12.6      Sample Size                                                                                                                                           42

12.7      Main Inclusion Criteria                                                                                                                    42

12.8      Main Exclusion Criteria                                                                                                                   43

12.9      Concomitant Medications                                                                                                                43

12.9.1       Prohibited Medications                                                                                                              43

12.9.2       Permitted Medications                                                                                                               44

12.10    Schedule                                                                                                                                                44

12.11    Outcome Measures (Per-Protocol)                                                                                                44

12.11.1     Primary Efficacy Measures                                                                                                        44

12.11.2     Secondary Efficacy Measures                                                                                                    44

12.11.3     Safety Measures                                                                                                                          44

12.11.4     Pharmacokinetic Measures                                                                                                        44

12.12    Analysis Plan (Per-Protocol)                                                                                                           45

12.12.1     General Considerations                                                                                                              45

12.12.2     Demographic And Baseline Characteristics                                                                             45

12.12.3     Study Hypotheses                                                                                                                        45

12.12.4     Primary Efficacy Parameters                                                                                                     45

12.12.5     Secondary Efficacy Parameters And Other Analyses                                                              45

12.12.6     Sample Size Calculation                                                                                                             46

12.12.7     Interim Analysis                                                                                                                          46

12.13    Protocol Amendments                                                                                                                       46

12.14    Post-Hoc Analysis Plan (Forest Laboratories)                                                                            46

12.14.1     Objectives                                                                                                                                    47

12.14.2     Efficacy Outcome Measures                                                                                                      47

12.14.3     Study Populations                                                                                                                        47

12.14.4     Patient Disposition And Study Completion                                                                              48

12.14.5     Demographic And Other Baseline Characteristics                                                                  48

12.14.6     Efficacy Analyses                                                                                                                        48

12.14.7     Exposure And Dosing Compliance                                                                                            49

12.14.8     Sample Size Estimate                                                                                                                  50

12.15    Key Changes Contained In Post-Hoc Analysis Plan                                                                    50

12.16    Efficacy Results                                                                                                                                   51

12.16.1     Patient Disposition                                                                                                                     51

12.16.2     Protocol Deviations                                                                                                                    51

12.16.3     Demographic And Other Baseline Characteristics                                                                  51

12.16.4     Brain Imaging At Study Entry                                                                                                     52

12.16.5     Extent Of Exposure And Compliance                                                                                        53

12.16.6     Primary Efficacy Analysis                                                                                                          53

12.16.7     “Primary Efficacy Analysis” On Dementia Of The Alzheimer’s Type Subset                       55

12.16.8     “Primary Efficacy Analysis” On Vascular Dementia Subset                                                   56

12.16.9     Analysis Of Secondary Efficacy Measures                                                                               58

12.16.10    Additional Analyses                                                                                                                    58

12.17    Sponsor’s Conclusions Regarding Efficacy                                                                                  59

12.18    Agency Statistical Review                                                                                                                60

12.19    Reviewer’s Comments                                                                                                                       60

13        Study MEM-MD-02                                                                           61

13.1      Study Protocol                                                                                                                                     61

13.1.1       Title                                                                                                                                              61

13.1.2       Objective                                                                                                                                      61

13.1.3       Design                                                                                                                                          61

13.1.4       Duration                                                                                                                                       61

13.1.5       Sample Size                                                                                                                                  61

13.1.6       Selection                                                                                                                                      61

13.1.7       Dosage                                                                                                                                          64

13.1.8       Schedule                                                                                                                                       64

13.1.9       Outcome Measures                                                                                                                     65

13.1.10     Safety Monitoring                                                                                                                       65

13.1.11     Statistical Analysis Plan                                                                                                             65

13.2      Efficacy Results                                                                                                                                   69

13.2.1       Patient Disposition                                                                                                                     69

13.2.2       Treatment Duration                                                                                                                     70

13.2.3       Dosing Compliance                                                                                                                     70

13.2.4       Demographic And Other Baseline  Characteristics                                                                 71

13.2.5       Primary Efficacy Analysis                                                                                                          72

13.2.6       Analysis Of Secondary Efficacy Measures                                                                               74

13.2.7       Additional Sponsor Analyses                                                                                                     76

13.2.8       Agency Subgroup Analysis                                                                                                         76

13.3      Sponsor’s Conclusions Regarding Efficacy                                                                                  77

13.4      Agency Statistical Reviewer’s Comments                                                                                    77

13.5      Reviewer’s Comments                                                                                                                       77

14        Additional Efficacy Studies                                                      77

14.1      Brief Outline Of Study Design                                                                                                         77

14.1.1       MRZ 9202                                                                                                                                   77

14.1.2       MRZ 9408                                                                                                                                   78

14.2      Efficacy Results                                                                                                                                   78

14.2.1       Patient Disposition                                                                                                                     79

14.2.2       Demographic And Other Baseline Characteristics                                                                  79

14.2.3       Results Of Analysis Of Primary Efficacy Parameters                                                             80

14.2.4       Subgroup Analysis Of ADAS-Cog                                                                                             80

14.3      Sponsor’s Conclusions                                                                                                                      81

14.4      Reviewer’s Comments                                                                                                                       81

15        Overall Comments About Key Efficacy Studies                  81

16        Recommendation                                                                          85

A. Oliva, M.D. ________                                                                             85


 

NOTE: The executive summary and main review contained in this document are intended only for briefing participants in a Peripheral and Central Nervous System Drugs (PCNS) Advisory Committee meeting that has been scheduled for September 24, 2003.   It is not the final version of the efficacy review of this application.

2        Executive Summary

This summary is restricted to an evaluation of the efficacy of memantine for the proposed indication. 

 

2.1       Recommendation

Deferred.

 

2.2       Proposed Indication

“The treatment of moderate-to-severe dementia of the Alzheimer’s type.”

 

2.3       Summary Of Clinical Findings

In support of the efficacy of memantine as a treatment for moderate-to-severe dementia of the Alzheimer’s type, the sponsor has submitted the results of 3 clinical studies.

 

  • A randomized, double-blind, placebo-controlled, parallel-arm study (MRZ 9605) of the efficacy of memantine in comparison with placebo in patients with moderate-to-severe dementia of the Alzheimer’s type.

 

  • A randomized, double-blind, placebo-controlled, parallel-arm study (MEM-MD-02) of the efficacy of memantine in comparison with placebo in patients with moderate-to-severe dementia of the Alzheimer’s type, already taking a stable dose of donepezil.

 

  • A randomized, double-blind, placebo-controlled, parallel-arm study (MRZ 9403) of the efficacy of memantine compared with placebo in patients with moderate-to-severe dementia of the Alzheimer’s, vascular, or mixed type.

 

These studies are summarized in greater detail below.

 

2.3.1     MRZ 9605

This study was conducted in the United States

 

2.3.1.1    Design

The two key criteria used for enrolling patients in this study were a diagnosis of probable Alzheimer’s Disease, using the National Institute for Neurological and Communicative Diseases and Stroke – Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) criteria, and a baseline Mini-Mental Status Examination score of 3 to 14. Patients taking acetylcholinesterase inhibitors or other drugs intended for the treatment of cognitive dysfunction were excluded from the trial.

 

Patients enrolled in this study were randomized to treatment with one of the following regimes for the  28-week period of double-blind, parallel-arm treatment

 

  • Memantine 10 mg b.i.d (reached by titration)

 

  • Placebo

 

The primary efficacy measures for the study a measure of function, a modification of the Alzheimer’s Disease Cooperative Study – Activities of Daily Living (ADCS-ADL) scale, and a global measure, the Clinician-Interview Based Impression of Change – Plus (CIBIC-Plus). Among the 7 secondary efficacy measures was the Severe Impairment Battery (SIB), a measure of cognition.

 

The primary efficacy analysis and the analysis of the secondary efficacy measures was carried out on an intent-to-treat (ITT) basis, using the last-observation-carried-forward (LOCF) method for imputing data; the statistical method used to compare the treatment groups was the Wilcoxon-Mann-Whitney test for independent samples.

 

2.3.1.2    Results

252 patients were enrolled in the study and randomized in exactly equal proportions to the 2 treatment groups. 97 memantine-treated patients and 84 placebo-treated patients completed the study.

 

Patients actually enrolled in this study had a baseline Mini-Mental Status Examination score that ranged from 1 to 14.

 

The primary efficacy analysis of the modified ADCS-ADL compared the mean change from baseline to endpoint between the memantine and placebo groups. While the difference between the treatment groups was small (2.00 points), it was statistically significant (p = 0.022) and in favor of memantine.

 

The primary efficacy analysis of the CIBIC-Plus compared the mean scores at endpoint between the memantine and placebo groups. Again, the difference between treatment groups was small (0.25 points) and did not quite reach pre-specified levels of statistical significance (p = 0.064), although the difference did favor memantine.

 

Analysis of the change from baseline to endpoint mean score on the SIB, using the LOCF method, yielded a nominally, but highly statistically significant p-value of 0.0003, for a mean group difference in score of 5.91 points that favored memantine.

 

2.3.2     MEM-MD-02

This study was conducted in the United States.

 

2.3.2.1    Design

The three key criteria used for enrolling patients in this study were a diagnosis of probable Alzheimer’s Disease, using the NINCDS-ADRDA criteria, a baseline Mini-Mental Status Examination score of 5 to 14, and treatment with donepezil for at least 6 months, with a stable dose for at least 3 months.

 

Patients enrolled in this study were randomized to treatment with one of the following regimes for the  24-week period of double-blind, parallel-arm treatment.

 

  • Memantine 10 mg b.i.d (reached by titration) plus donepezil

 

  • Placebo plus donepezil

 

The primary efficacy measures for the study consisted of a subset of the Alzheimer’s Disease Cooperative Study – Activities of Daily Living (ADCS-ADL) scale, as a measure of function,  and the Severe Impairment Battery (SIB), as a measure of cognition. The study also had multiple secondary efficacy measures.

 

The primary efficacy analysis was carried out on an intent-to-treat (ITT) basis, using the last-observation-carried-forward (LOCF) method for imputing data; the statistical method used to compare the treatment groups was a two-way analysis of covariance (ANCOVA), with treatment group and center as main effects and baseline score as the covariate.

 

2.3.2.2    Results

404 patients were enrolled in the study. They were randomized as follows to the 2 treatment groups.

 

  • Memantine plus donepezil: 203 patients        
  • Placebo plus donepezil: 201 patients

 

Patients actually enrolled in this study had a baseline Mini-Mental Status Examination score that ranged from 5 to 16.

 

322 patients completed the study. Their distribution among the treatment groups was as follows:

 

  • Memantine plus donepezil: 172 patients        
  • Placebo plus donepezil: 150 patients

 

The primary efficacy analysis of the modified ADCS-ADL compared the mean change from baseline to endpoint between the memantine plus donepezil and placebo plus donepezil groups. Although the difference between the treatment groups was small (1.40 points), it was statistically significant (p = 0.028) and favored memantine.

 

The primary efficacy analysis of the SIB also compared the mean change from baseline to endpoint between the 2 treatment groups. While small (3.40 points), the treatment difference between the groups was statistically significant (p < 0.001) and favored memantine.

 

2.3.3     MRZ 9403

This study was conducted in Latvia.

 

2.3.3.1    Design

Key inclusion criteria for this study were the presence of dementia, according to DSM-III-R, a Mini-Mental Status Examination score of < 10, and Global Deterioration Scale staging of 5 to 7; the dementia could of the Alzheimer’s, vascular, or mixed variety, without any diagnostic criteria for these conditions being specified. Those enrolled in the study were then classified, after their enrollment in the study, and based on their Hachinski Ischemic Scale score, as having either vascular dementia or Alzheimer’s Disease.

 

Patients enrolled in this study were randomized to treatment with one of the following regimes for the  12-week period of double-blind, parallel-arm treatment.

 

  • Memantine 10 mg q.d. (reached by titration)

 

  • Placebo

 

The protocol-designated primary efficacy measures were as follows

 

  • The Behavioral Rating Scale in Geriatric Patients (BGP) Care Dependency Subscale, a measure of activities of daily living and behavior. This is in turn a subset of the BGP proper.

 

  • The Clinician Global Impression of Change (CGI-C), a global measure. For use as a primary efficacy measure, the original 7-point scale was to be dichotomized.

 

A third primary efficacy measure was introduced when a second analysis plan was formulated several years after the study blind was broken, and the study results published. This measure, designated as the BGP Cognitive Subscale was an ad-hoc subset of the BGP Care Dependency Subscale, and contained 5 items (that were considered to measure cognition) out of 23 items in the BGP Care Dependency Subscale.

 

When the post-hoc statistical analysis plan was formulated, the original 7-point CGI-C was designated as a primary efficacy measure, instead of the dichotomized scale.

 

The protocol-specified primary efficacy analysis was to be done on the intent-to-treat population. As part of this analysis, the treatment groups were to be compared on the change from baseline score for the BGP Care Dependency Subscale using Wilcoxon-Mann-Whitney U tests. Analysis of the CGI-C (dichotomized) was to be carried out using Fisher’s exact test. Missing data were to be imputed using the worst possible score (worst change) for each efficacy parameter.

 

In the statistical analysis plan formulated post-hoc, the analysis of all 3 primary efficacy measures was to be based on the Wilcoxon rank-sum test, stratified by center.

 

2.3.3.2    Results

166 patients were enrolled in the study and randomized as follows:

 

  • Memantine: 82 patients

 

  • Placebo: 84 patients

 

158 patients completed the study and were distributed as follows:

 

  • Memantine: 78 patients

 

  • Placebo: 80 patients

 

The results of the protocol-specified primary efficacy analysis were as follows:

 

  • 73.2% of memantine-treated patients versus 45.2% of placebo-treated patients were considered responders at endpoint on the dichotomized CGI-C; the difference was statistically significant (p < 0.001).

 

  • The difference between the 2 treatment groups on the mean change from baseline to endpoint on the BGP Care Dependency Subscale score was 1.9 in favor of memantine (p = 0.016).

 

The results of the post-hoc primary efficacy analysis were as follows:

 

  • The difference between the treatment groups on the mean CGI-C score (7-point scale) at endpoint was 0.4 and in favor of memantine (p < 0.001).

 

  • The difference between the 2 treatment groups on the mean change from baseline to endpoint on the BGP Care Dependency Subscale score was 2.0, and in favor of memantine (p = 0.012).

 

  • The difference between the 2 treatment groups on the mean change from baseline to endpoint on the BGP Cognitive Subscale score was 0.8 and in favor of memantine (p = 0.001).
2.3.3.2.1     Subset Analysis

Patients who were enrolled in the study and randomized were classified after enrollment as having either dementia of the Alzheimer’s type or vascular dementia based on their modified Hachinski Ischemic Scale at study entry (they were considered to have dementia of the Alzheimer’s type if their score was £ 4).

 

79 patients subsequently diagnosed to have dementia of the Alzheimer’s type entered the study. Their distribution among the treatment groups was as follows

 

  • Memantine: 41 patients

 

  • Placebo: 38 patients

 

76 patients diagnosed to have dementia of the Alzheimer’s type completed the study and were distributed as follows:

 

  • Memantine: 39 patients

 

  • Placebo: 37 patients

 

The results of the analysis of the dementia of the Alzheimer’s type subset, using the same methods as used for the post-hoc primary efficacy analysis, were as follows:

 

  • The difference between the treatment groups on the mean CGI-C score (7-point scale) at endpoint was 0.4 and in favor of memantine (p = 0.003).

 

  • The difference between the 2 treatment groups on the mean change from baseline to endpoint on the BGP Care Dependency Subscale score was 3.0, and in favor of memantine (p = 0.003).

 

  • The difference between the 2 treatment groups on the mean change from baseline to endpoint on the BGP Cognitive Subscale score was 1.0 and in favor of memantine (p = 0.007).

 

A similar analysis performed on the vascular dementia subset, revealed statistically significant differences (p < 0.05) favoring memantine only for the CGI-C (7-point scale).

 

2.4       Conclusions

  • Based on the paradigm used for demonstrating the efficacy of drugs intended for the treatment of mild-to-moderate Alzheimer’s Disease, it appears appropriate that a claim for memantine in the treatment of moderate-to-severe Alzheimer’s Disease should be supported by evidence of efficacy on both a cognitive efficacy measure and, separately, on a global or functional primary efficacy measure

 

  • On the above basis, Studies MRZ 9605 and MEM-MD-02 have provided sufficient evidence to support the efficacy of memantine in moderate-to-severe dementia of the Alzheimer’s type. This evidence is as follows:

 

    • Patients enrolled in both studies had probable Alzheimer’s Disease and a baseline Mini-Mental Status Examination score that ranged from 1 to 16.

 

    • In Study MRZ 9605, a study evaluating memantine as monotherapy in a dose of 10 mg b.i.d, evidence for efficacy was seen on the Severe Impairment Battery, a measure of cognition, and on the modified ADCS-ADL scale, a measure of activities of daily living.

 

    • In Study MEM-MD-02, a study evaluating the efficacy of memantine, in a dose of 10 mg b.i.d as add-on therapy in patients already taking a stable dose of donepezil, evidence for efficacy was again seen on the Severe Impairment Battery and modified ADCS-ADL

 

    • The Severe Impairment Battery and modified ADCS-ADL, have at the very least, face validity as measures that can be used in patients with moderate-to-severe cognitive impairment

 

  • Study MRZ 9403 provides less-than-convincing evidence of the efficacy of memantine in moderate-to-severe dementia of the Alzheimer’s type

 

    • Patients enrolled in this study could have Alzheimer’s Disease, vascular dementia, or mixed dementia

 

    • 48% of patients enrolled in this study did not undergo brain imaging of any kind.

 

    • This study lacked a satisfactory cognitive outcome measure, and especially one that was prospectively-designated

3        Introduction

This submission contains a New Drug Application (NDA) for memantine hydrochloride tablets, which the sponsor is seeking to market for the treatment of moderate-to-severe dementia of the Alzheimer’s type. 

 

This review also evaluates data contained in an Amendment to this NDA, which was submitted on 1/10/03.

 

The efficacy of memantine for the proposed indication is considered by the sponsor to be based on 3 pivotal efficacy studies contained in this application. The reports of 2 of these studies (MRZ 90001-9605 and MRZ 90001-9403) are contained in the original application. The report of a third study (MEM-MD-02) constitutes most of the Amendment submitted on 1/10/03.

 

This submission is confined to reviewing data that are intended to support the efficacy of memantine. Data contained in this submission that are intended to support the safety of memantine are being reviewed separately by Dr Gerald Boehm of this Division.

 

The statistical review of efficacy data contained in this submission is being performed by Dr Tristan Massie.

 

Memantine has been developed for the treatment of moderate-to-severe dementia of the Alzheimer’s type under IND 33392. The previous sponsor of that application was Merz and Company, with whom this Division earlier had a number of discussions about the development of this drug.

 

In this review , the terms “dementia of the Alzheimer’s type” and “Alzheimer’s Disease” are often used interchangeably.

4        Organization Of Submission

The original submission of this NDA consists of 437 print volumes; the Amendment of 1/10/03 consists 24 print volumes. Selected components of the print application are also provided in electronic format; Case Report Forms and Case Report Tabulations (SAS transport files) are provided electronically only.

 

The reports of the efficacy studies that are considered pivotal are contained in the following print volumes.

 

Study

Volume

MRZ 9605

Volumes 117 – 141 of original application

MRZ 9403

Volumes 142 – 145 of original application

MEM-MD-02

Volumes 1 – 23 of Amendment (submitted Jan 10, 2003)

 

An Integrated Summary of Effectiveness is contained in Volumes 263 – 264 of the original application. In addition to summarizing the results of the 2 pivotal efficacy studies contained in that submission (MRZ 9605 and MRZ 9403), the Integrated Summary of Efficacy also summarizes data from 2 efficacy studies of memantine in mild-to-moderate probable vascular dementia (MRZ 9202 and MRZ 9408) the results of which, the sponsor believes, are pertinent to the claim that the sponsor is currently seeking.

5        Outline Of Review

This review will address the 3 pivotal efficacy studies, using information contained in the respective study reports; this will be supplemented by information contained in the Integrated Summary of Effectiveness, ancillary study reports, and related electronic components. The review will consist of the following in  the same order as below:

 

·       Chemistry

·       Proposed mechanism of action

·       Summary of memantine pharmacokinetics

·       Rating scales/outcome measures used in the key efficacy studies

·       Summary of the key efficacy studies (in table form)

·       Review of the key efficacy studies individually

·       Summary of additional efficacy studies

·       Overall comments about efficacy of memantine for the proposed indication

·       Review of draft labeling

·       Site inspection summary

·       Financial disclosure certification

·       Recommendations

6        Chemistry

The chemical name for memantine hydrochloride is 1-amino-3,5,-dimethyladamantane hydrochloride. The chemical structure of memantine is in the following figure.

 

 

The sponsor has proposed that memantine be marketed for oral administration as capsule-shaped film-coated tablets, containing the equivalent of 5 mg, 10 mg, 15 mg, and 20 mg of memantine hydrochloride.

 

Please see the Agency Chemistry review for further details.

7        Proposed Mechanism Of Action

The sponsor has suggested that memantine exerts its effects in Alzheimer’s Disease as follows:

 

  • Memantine is a moderate-affinity, uncompetitive, N-methyl-D-aspartate (NMDA) receptor antagonist that binds preferentially to the NMDA receptor-operated cation channel.
  • The NMDA receptor is activated by glutamate. Glutamate neurotoxicity may have a role in the pathogenesis of Alzheimer’s Disease.
  • Non-clinical evidence suggests that blockade of NMDA receptors by memantine can provide protection from the neurotoxic effects of glutamate, and improve memory and learning.

 

For further details, please see the Agency Pharmacology review.

8        Summary Of Memantine Pharmacokinetics And Clinical Pharmacology

The following is based on information provided by the sponsor in the Application Summary.

 

  • Following oral administration, memantine is completely absorbed, with a tmax of 4 to 6 hours, and an oral bioavailability of 100%.
  • Food does not affect the bioavailability of memantine administered as a tablet.
  • Exposure levels, based on Cmax and AUC0-∞, are dose-proportional after single doses ranging from 10 to 40 mg.
  • Memantine is extensively distributed in tissues and readily crosses the blood-brain barrier.
  • Memantine is about 45% protein-bound.
  • The terminal half-life of memantine is 60 to 80 hours with no changes in half-life over the 5 to 40 mg single-dose range.
  • Memantine undergoes little metabolism and is excreted largely (75 to 90%) unchanged in the urine (and in part by renal tubular secretion); the remainder is converted to 3 polar metabolites - the N-gludantan conjugate, 6-hydroxy memantine, and 1-nitroso-deaminated memantine - all of which have minimal or no NMDA receptor antagonist activity.
  • Memantine clearance is reduced with increasing degrees of renal impairment. No dosage adjustment, based on age and gender, is felt to be needed.
  • The CYP450 system is minimally involved in the metabolism of memantine. Based on in-vitro studies, memantine produces only minimal inhibition of CYP450 isoenzymes CYP1A2, CYP2A6, CYP2C9, CYP2D6, CYP2E1, and CYP3A4.
  • Memantine does not have any pharmacokinetic or pharmacodynamic interaction with donepezil.

 

Please see the Agency Biopharmaceutics and Clinical Pharmacology review for further details.

9        Rating Scales/Outcome Measures Used In Key Efficacy Studies

In this section I will summarize instruments used as primary and secondary efficacy measures for key studies included in this application, as well as those used to evaluate patients at the time of entry into these studies

9.1       Primary Efficacy Variables

9.1.1     Severe Impairment Battery

This scale has been developed to assess cognitive function in severely demented patients. It is divided into 9 sub-scales assessing attention, orientation, language, memory, praxis, visuospatial perception, construction, social skills and orientation to name. The tests that comprise the Severe Impairment Battery involve simple 1-step commands that may be presented with gestural cues; 51 such tests are assessed altogether. Total scores range from 0 to 100 points with higher scores indicating better cognitive function.

 

The test-retest reliability, construct validity and sensitivity to change of the Severe Impairment Battery have been evaluated (Schmitt FA et al. The severe impairment battery: concurrent validity and the assessment of longitudinal change in Alzheimer’s disease. The Alzheimer’s Disease Cooperative Study. Alzheimer Dis Assoc Disord. 1997;11 Suppl 2:S51-6) in a one-year study. The results may be summarized as follows

 

·        Test-retest reliability was assessed using baseline to one-month, and baseline to two-month correlations in 90 patients. Correlations were statistically significant for the following Mini Mental Status Examination score groups at one month: 0-4, 5-9 and 10-15, but not for the 16-20 group. At 2 months correlations were seen for all groups

·        Construct validity was assessed by comparison with the following: CDR, CDR “sum of boxes”, FAST, GDS and Mini Mental Status Examination. Baseline scores were compared on 192 patients. Statistically significant correlations were demonstrated between the Severe Impairment Battery and each of the other measures


·        Sensitivity to change was assessed using in comparison with CDR, CDR “sum of boxes”, FAST, and GDS. 180 patients were evaluated over one year. Correlations were best for subjects with baseline Mini Mental Status Examination scores in the 5-9 range as indicated by the following table.

 

9.1.2     Alzheimer's Disease Cooperative Study-Activities of Daily Living (ADCS-ADL): Modified

This is a rating scale used to assess basic and instrumental activities of daily living. In the full version of the scale, 45 items are rated by the investigator using information supplied by the caregiver. Each item has a score range varying from 0-3 to 0-7. Higher scores indicate better function.

 

In the studies described below, a modified version of the ADCS-ADL was used consisting of a subset of 19 of the above 45 items. These 19 items, selected to fit the expected activities of daily living profile of patients with moderate-to-severe dementia, consist of the following:

 

Eating

Ability to watch TV

Ability to be left alone

Walking

Making conversation

Ability to turn a faucet on

Toileting

Clearing a table

Ability to turn a faucet off

Bathing

Locating belongings

Ability to turn a light on

Grooming

Obtaining a beverage

Ability to turn a light off

Dressing

Litter disposal

 

Use of a telephone

Traveling outside the home

 

 

For the modified ADCS-ADL, a sum score was calculated by adding the scores for the individual items, and used as a primary efficacy measure. The sum score could range from 0 to 54, with higher scores indicating better function.

 

A second method of scoring the modified ADCS-ADL items has been used to derive a  secondary efficacy measure. Each post-baseline item score has been divided into 2 categories, and each category rated as follows

 

Unchanged or improved score:

Rated as an improvement

Declining score:

Rated as a deterioration

 

The sum of the scores for those items rated as an improvement was used as the secondary efficacy measure.

9.1.3     Clinician Interview Based Impression of Change-Plus (CIBIC-Plus)

The format for this instrument consists of the assessment of an independent clinician based on observation of the patient at an interview, and information provided by the caregiver. The clinician is blinded to the results of other study assessments. The clinician’s overall impression of the global change in disease severity, compared with baseline, is rated.  A 7-point categorical rating scale is used, ranging from a score of 1 indicating “markedly improved”, to a score of 7 indicating “markedly worse”, and with a score of 4 indicating “no change”.

 

The CIBIC-Plus was also a secondary efficacy measure in a study.

9.1.4     Clinician Global Impression of Change (CGI-C)

This instrument was used in a single study. The format for this instrument in that study was similar to the CIBIC-Plus except that the rater had access to all information (including psychometric scores and physical examination results) at baseline, when the severity of the disease (Clinical Global Impression of Severity [CGI-S]) was assessed. Subsequent ratings were based only on patient assessment and on information provided by the caregiver.

 

The CGI-C was scored using the same 7-point scale that was used for the CIBIC-Plus.

 

Analyses of the CGI-C used either the original 7-point scale, or a dichotomized scale; the dichotomized scale grouped patients into responders (CGI-C scores of 1 to 4) and non-responders (CGI-C score of 5 to 7)

9.1.5     Behavioral Rating Scale In Geriatric Patients (BGP)

The BGP itself is a 35-item clinician-rated measure that assesses behavior (including mood), basic cognitive functions, mobility and activities of daily living. Each item is rated from 0-2, with 2 indicating the worst level of functioning. For example the item “requires assistance with eating” is rated as follows: 0 = no assistance; 1 = limited assistance and 2 = frequently. Rating is based upon direct observation by the clinician

 

The BGP has 4 standard subscales

 

·       Care Dependency Subscale

·       Aggressiveness Subscale

·       A composite subscale comprising physical disability, depression, and mental disability items

·       Inactivity Subscale

 

The BGP Care Dependency Subscale comprises 23 out of the 35 items in the entire BGP. The items assessed by this subscale are representative of either activities of daily living or behavior. Each item is scored on a scale from 0 to 2. The maximum score on this sub-scale is 46, with higher scores indicating a worse level of function.

 

An ad-hoc (and post-hoc) subscale derived from the BGP, termed the BGP Cognitive Subscale, was used in the key efficacy study 9403. This subscale comprised 5 out of 23 items in the BGP Care Dependency Subscale. Each item was rated on a scale from 0 to 2. The maximum score for this subscale was10 with a higher score indicating a worse level of functioning. The items that were rated as part of the BGP Cognitive Subscale were as follows:

 

Item

Scoring

Cognitive Domain Assessed (According To Sponsor)

The patient makes himself understood

(by speaking, writing, or gestures)

Always

0

Expressive speech

Sometimes

1

Rarely

2

 

The patient finds his way in the nursing home (e.g., to his room, to the toilet, to his place at the table)

Generally yes

0

Spatial orientation

Some ways yes, others no

1

Generally no

2

 

The patient understands in what home or clinic he is

Always

0

Orientation for place

Sometimes

1

Rarely

2

 

The patient knows the names of the stuff (sic)

More than one

0

Naming

Only one

1

None

2

 

The patient understands what you communicate with him (by speaking, writing, or gestures)

Always

0

Receptive language function

Sometimes

1

Never

2

 

9.2       Secondary Efficacy Variables

9.2.1     Mini-Mental Status Examination (MMSE)

This is a multi-item instrument that examines orientation, registration, attention, calculation, recall, visuospatial abilities and language. The maximum score is 30, with higher scores indicating better cognitive function.

9.2.2     Functional Assessment Staging

This instrument is intended to assess functional decline in patients with Alzheimer’s Disease. It evaluates a patient’s ability to perform a variety of functions. The scale has seven major stages ranging from Stage I (“normal”) to Stage 7 (“severe”); Stage 6 is further divided into 5 subsets (6a to 6e); and Stage 7 is further divided into 6 subsets (7a to 7f). Staging is based on specific deficits in functional ability

9.2.3     Neuropsychiatry Inventory

This is a validated instrument that assesses the following 10 domains (subscales): delusions, hallucinations, agitation/aggression, depression/dysphoria, anxiety, elation/euphoria, apathy/indifference, disinhibition, irritability/lability and aberrant motor behavior. Each item is rated according to its frequency and severity; rating is based on interviewing a caregiver. The maximum total score (the sum of the subscale scores) is 120 with a higher score indicating greater behavioral abnormality.

9.2.4     Resource Utilization In Dementia

This instrument is designed to assess caregiver burden for those caring for patients with Alzheimer’s Disease. The assessment consists of a structured interview with the caregiver and has 2 parts

 

Part A: This is a questionnaire administered at baseline

Part B: This is a follow-up questionnaire

 

The questionnaires assess basic demographic information, significant health events since the first questionnaire was administered, time spent with patient, changes in caregiver’s work status and changes in health care utilization

9.2.5     G2 Scale

This is a 16-item nurse-rated scale that assesses the following: cognition, mobility, behavior, and activities of daily living. The scale is rated in 2 ways

9.2.5.1    G2 Condition (G2)

In this method of rating, patient evaluations at specific timepoints are independent of each other. Each item is rated on a 6-point categorical scale with a higher score indicating more severe impairment

9.2.5.2    G2 Change (G2-C)

In this method of rating, the patient’s condition at specific timepoints is rated in comparison with baseline on a 7-point categorical scale ranging from 1 (“very much improved”) to 7 (“very much worse”)

9.2.6     Instrumental Activities of Daily Living Performance Test  (IADLPT)

This is a nurse-rated measure that evaluates a patient’s ability to perform specific motor activities of daily living. The activities assessed are as follows: buttoning and unbuttoning 3 buttons; opening and closing 3 safety pins; making a knot and bow with a shoelace; applying a plaster (bandage); and reading and dialing a 6-digit phone number. Each activity is rated based on time taken, and on quality (1 = good; 2 = moderate; and 3 = bad)

9.2.7     Clinical Global Impression of Severity (CGI-S)

The severity of Alzheimer’s Disease was graded according to the following scale in an efficacy study included in this application.

Score

Severity of disease

1

Normal, not at all ill

2

Borderline mentally ill

3

Mildly ill

4

Moderately ill

5

Markedly ill

6

Severely ill

7

Among the most extremely ill

9.2.8     Clinician Global Impression of Change (CGI-C) Benefit/Risk Index

This measure was the ratio of the CGI-C Efficacy (Benefit) Index to the CGI-C Risk Index

 

The CGI-C Efficacy Index was rated based on a 4-point scale that ranged from 1 to 4 (1 to 3 for good to minimal improvement; 4 for unchanged or worse)

 

For the CGI-C Risk Index, adverse events were rated according to the following 4-point scale:

 

No adverse events:

1

No significant interference with function:

2

Significant interference with function:

3

Adverse events outweigh therapeutic benefits:

4

 

9.3       Rating Scales Not Used As Efficacy Variables

9.3.1     Hamilton Depression Scale (HDS)

This is an observer-rated measure that is used to assess the severity of depression based on an interview of the patient and caregiver. 21 symptoms (e.g., anxiety, feelings of guilt, depressed mood) are each rated based on a structured categorical scale with a higher score indicative of a greater severity of symptoms. Nine of the items are rated on a 5-point scale (0 to 4), 11 items are rated on a 3-point scale (0 to 2), and a single item on a 4-point scale (0 to 3)

9.3.2     Global Deterioration Scale (GDS)

This is an instrument intended to assess the magnitude of cognitive, functional and behavioral decline. A clinician provides an overall rating for the patient on a scale from 1, indicating “no cognitive decline”, to 7, indicating “very severe cognitive decline” as in the table below. Guidelines for rating the individual for each integral value on the scale from 1 through 7 are specified.

Stage

Stage

1

No cognitive decline

2

Very mild cognitive decline

3

Mild cognitive decline

4

Moderate cognitive decline

5

Moderately severe cognitive decline

6

Severe cognitive decline

7

Very severe cognitive decline

9.3.3     Hachinski Ischemic Scale (Rosen Modification)

This is a nine-item instrument that is intended to help distinguish between vascular dementia and Alzheimer’s Disease. The items assessed consist of the following: abrupt onset; stepwise deterioration; fluctuating course; somatic complaints; emotional incontinence; history of hypertension; history of stroke; focal neurological symptoms; and focal neurological signs. Each of the items is assigned a pre-specified score of either  “1” or “2” if present; items rating a score of “2” are abrupt onset, fluctuating course, history of stroke, focal neurological symptoms and focal neurological signs. The maximum score is 14 with higher scores being considered more indicative of vascular dementia.

10    Tabular Summary Of Key Efficacy Studies

The sponsor has submitted the reports of 3 studies that are intended to support the claim for memantine in the treatment of moderate-to-severe Alzheimer’s Disease. These studies are outlined below. For full details about each of these studies, please see the individual study summaries later in the review.

10.1   Study MRZ 9605

This study was performed in the United States under IND 33392, and is outlined in the table below.

 

Design

Randomized, double-blind, placebo-controlled, parallel-group

Duration

28 weeks of double-blind, parallel-arm treatment

Key Inclusion Criteria

Probable Alzheimer’s Disease

Mini-Mental Status Examination: 3-14

GDS: Stages 5-6

FAST ³ 6a

Primary Outcome Measures

ADCS-ADL, CIBIC-Plus

Population For Primary Efficacy Analysis

Intent-to-treat-LOCF

Secondary Outcome Measure

SIB, NPI, Global Deterioration Scale, Categorical ADCS-ADL, Functional Assessment Scale, Resource Utilization in Dementia

Dose Arms

Memantine 10 mg b.i.d

Placebo

Number randomized

126

126

Number completing

97

84

Note that the mean Mini-Mental Status Examination at study entry was 7.9

 

The results of this study are summarized in the table below

 

 

                      LOCF Analysis

                     OC Analysis

Memantine

(n = 126)

Placebo

(n = 126)

p-value*

 

Memantine

(n = 97)

Placebo

(n = 84)

p-value*

 

CIBIC-Plus

4.48

4.73

0.064

4.38

4.74

0.025

ADCS-ADL

-3.02

-5.02

0.022

-2.49

-5.48

0.003

SIB

-3.93

-9.84

< 0.001

-4.46

-10.16

0.002

*p-values are based on Wilcoxon-Mann-Whitney test for between treatment comparisons

10.2   Study MRZ  9403

This ex-IND study was conducted in Latvia

Design

Randomized, double-blind, placebo-controlled, parallel-group

Duration

12 weeks of double-blind, parallel-arm treatment

Key Inclusion Criteria

Alzheimer’s Disease, vascular dementia, or mixed dementia*

Mini-Mental Status Examination: 0-9

GDS: Stages 5-7

Primary Efficacy Measures

BGP Care Dependency Subscale

CGI-C

Secondary Efficacy Measures

G2, G2-C, IADL

Post-Hoc Primary Efficacy Measure

BGP Cognitive Subscale

Dose Arms

Memantine 10 mg daily

Placebo

Number randomized

82

84

Number completing

78

80

*Randomization was not stratified by dementia type. Using the Hachinski Ischemic Scale, all patients enrolled in the study were grouped post-hoc into 2 categories: Alzheimer’s Disease and vascular dementia

 

Only a total of 86 patients (40 placebo and 46 memantine) had brain imaging studies (CT scan only) done

 

The results of the post-hoc primary efficacy analysis for this study are summarized in the table below

 

                      LOCF Analysis

                     OC Analysis

Memantine

(n = 82)

Placebo

(n = 84)

p-value*

 

Memantine

(n = 78)

Placebo

(n = 80)

p-value*

 

CGI-C

3.09

3.52

0.001

3.01

3.48

0.001

BGP Care Dependency

-5.29

-3.27

0.012

-5.56

-3.50

0.010

BGP Cognitive

-1.85

-1.12

0.001

-1.95

-1.19

0.001

*p-values are based on Cochran-Mantel-Haenszel test for row means (using modified ridit score) controlling for center


10.3   Study MEM-MD-02

This study was conducted in the United States under IND 33392, and is outlined in the table below

 

Design

Randomized, double-blind, placebo-controlled, parallel-group

Duration

24 weeks of double-blind, parallel-arm treatment

Key Inclusion Criteria

·        Probable Alzheimer’s Disease

·        Mini-Mental Status Examination: 5-14

·        Treatment with donepezil for at least 6 months, and on a stable dose for 3 months

Primary Outcome Measures

·        Severe Impairment Battery

·        ADCS-ADL (modified)

 

Population For Primary Efficacy Analysis

Intent-to-treat-LOCF

Secondary Outcome Measure

·        CIBIC-Plus

·        Neuropsychiatry Inventory

·        Functional Assessment Staging

·        Resource Utilization In Dementia

·        Behavioral Rating Scale For Geriatric Patients

 

Dose Arms

Memantine 10 mg b.i.d  + donepezil

Placebo + donepezil

Number randomized

203

201

Number completing

172

150

Note that the mean Mini-Mental Status Examination (± standard deviation ) at study entry was 9.9 (3.13) in the memantine  plus donepezil group and 10.2 (2.98) in the placebo plus donepezil group

 

The results of the primary efficacy analysis for this study are summarized in the tables below

10.3.1  Least Square Mean Change From Baseline In Severe Impairment Battery

 

 Placebo/Donepezil

 Memantine/Donepezil

p-value

N

Mean

N

Mean

Week 24 (LOCF)

196

-2.5

198

0.9

< 0.001

Week 24 (OC)

153

-2.4

171

1.0

< 0.001

10.3.2  Least Square Mean Change From Baseline In ADCS-ADL 

 

 Placebo/Donepezil

 Memantine/Donepezil

p-value

N

Mean

N

Mean

Week 24 (LOCF)

197

-3.4

198

-2.0

0.028

Week 24 (OC)

152

-3.3

172

-1.7

0.020

 


11    Study MRZ 9605

This study was conducted at 32 centers in the United States

11.1   Study Protocol

The version of the protocol summarized below is the final one, and does not appear to have been amended further before the study blind was broken.

11.1.1  Objective   

To demonstrate that memantine is superior to placebo, as assessed by global and functional measures, in treating patients with moderately severe Alzheimer's Disease.

11.1.2  Design   

Randomized, double-blind, placebo-controlled, parallel-arm trial of 28 weeks duration.

 

The proposed study was to be followed by an optional 24-week open-label period during which all patients were to receive the active drug

11.1.3  Sample Size  

250 patients randomized equally to the 2 treatment groups

11.1.4  Key Inclusion Criteria

·       Men or post-menopausal/surgically sterile women > 50 years old

·       Probable Alzheimer's Disease, according to DSM-IV and NINCDS-ADRDA criteria

·       Clinical and psychometric rating scores as follows:

·        Mini-Mental Status Examination range of 3-14

·        Global Deterioration Scale 5 or 6

·        Functional Assessment Scale Score ³ 6a

·        Hachinski Ischemic Scale score (as modified by Rosen) £ 4

·       CT or MRI of brain, within 12 months prior to randomization, compatible with Alzheimer's Disease

·       Ability to walk, at least with an assistive device

·       Vision and hearing sufficient to comply with testing

·       Normal cognitive and social functioning prior to onset of dementia

·       Consistent caregiver to accompany patient to assessment visits as far as possible

·       Sufficient basic education to be testable

·       Living outside an institution

·       Informed consent from patient, caregiver, legal guardian (if applicable) and a witness

11.1.5  Key Exclusion Criteria

·       Dementia to any condition other than Alzheimer's Disease, including vascular dementia (modified Hachinski Ischemic Scale ³ 5; positive NINDS-AIREN criteria)

·       Significant neurological disease other than Alzheimer’s Disease, including cerebral tumor, Huntington’s Disease, Parkinson’s Disease, normal pressure hydrocephalus, and other entities

·       Major depression according to DSM-IV

·       Psychotic episodes requiring hospitalization or antipsychotic therapy for more than 2 weeks within the past 10 years, not linked to Alzheimer’s Disease

·       Agitation sufficient to preclude participation in this trial

·       Alcohol or drug dependence diagnosed within the past 10 years

·       Epilepsy or anti-epileptic drug therapy

·       Abnormal laboratory tests that might point to another etiology for dementia: serum B12, folate, thyroid functions, electrolytes, syphilis serology

·       Musculoskeletal diseases that could interfere with assessment

·       Acute or poorly controlled medical illness: blood pressure > 180 mmHg  systolic or 100 mmHg diastolic; myocardial infarction within 6 months; uncompensated congestive heart failure (NYHA Class III or IV), severe renal, hepatic or gastrointestinal disease that could alter drug pharmacokinetics; blood glucose > 180 mg/dl on repeated testing at entry into study or need for insulin therapy

·       Previous randomization in this trial or participation in another investigational trial < 2 months prior to randomization

·       Likelihood, according to clinical judgement, of being transferred to a nursing home within 6 months

11.1.6  Concomitant Medications

11.1.6.1                Prohibited Medications:

Investigational drugs, anticonvulsants, anti-Parkinsonian drugs, benzodiazepines, barbiturates, other hypnotics, neuroleptics, initiation of antidepressant and anxiolytic medication,  cholinesterase inhibitors (the last of these may be used after a 30 day washout period), other drugs intended for the treatment of cognitive dysfunction

11.1.6.2                Permitted medications: 

Chloral hydrate as a hypnotic (not within 24 hours of an assessment; maximum dose 2000 mg/day), xanthine derivatives (if dose remains stable throughout trial), beta-blockers and estrogens (if dose remains stable for 3 months prior to or during trial), “anti-inflammatory” drugs (if dose is constant for at least 1 month before trial, unless drug is used on an acute basis in which case the drug should not be used except for 3 days prior to each assessment), Ginkgo (if not investigational), Vitamin E and coenzyme Q (if dose is constant for at least 1 month before trial), all other medications (without restrictions)

11.1.7  Dosage  

The dosing regime for the double-blind phase study is summarized in the following table

 

  

 

Matching placebo was to be used as indicated above, during the double-blind phase.

11.1.8  Duration

28 weeks of double-blind treatment


 

11.1.9  Schedule  

The study schedule is summarized in the following table which I have copied from the submission

11.1.10                  Outcome Measures

11.1.10.1            Primary Efficacy Measures  

·       Clinician Interview Based Impression of Change-Plus

·       Alzheimer’s Disease Cooperative Study-Activities of Daily Living (modified inventory)

11.1.10.2            Secondary Efficacy Measures  

·       Functional Assessment Scale

·       Mini-Mental Status Examination

·       Severe Impairment Battery

·       Global Deterioration Scale

·       Modified ADCS-ADL: Sum Scores of Responses

·       Neuropsychiatry Inventory: Total Score (based on frequency and severity of each behavior) and NPI Caregiver Distress Scale

·       Resource Utilization in Dementia

11.1.10.3            Safety Variables

Adverse events, vital signs, laboratory tests, electrocardiograms

11.1.10.4            Pharmacokinetic Measures

Plasma level of memantine

11.2   Analysis Plan

The analysis plan, finalized 11/29/99 after discussions with the Division, will be reviewed only as it pertains to the assessment of efficacy

11.2.1  General Considerations

·       All statistical tests on the primary and secondary efficacy variables were to be 2-sided and a p-value of < 0.05 was to be considered statistically significant

11.2.2  Study Populations

·       The intention-to-treat population was to consist of every patient randomized regardless of whether the patient received any treatment at all or the correct treatment.

·       The treated-per-protocol population was to consist of the intention-to-treat population excluding patients with any of the following: no measurement of primary efficacy variables after 28 weeks of treatment; intake of less than 75 % of the prescribed individual daily dose in the course of the trial; major deviations from the protocol; violation of inclusion or exclusion criteria and change in caregiver status without adequate substitution or supervision.

·       The evaluable-for-safety population was to consist of all those randomized who received at least one dose of study medication 

·       Retrieved dropout analyses were also planned for those patients missing Week 28 data

11.2.3  Demographic And Baseline Characteristics

The analysis plan does not specifically state how these parameters were to be analyzed

11.2.4  Drug Compliance

Overall compliance for the study was to be computed as follows: 100 x [(total number of tablets dispensed) – (total number of tablets returned) – (total number of tablets reported lost)]/[(2 x number of days for which 2 tablets were prescribed per day plus number of days for which 1 tablet was prescribed per day)]

11.2.5  Primary Efficacy Parameters

·       The primary efficacy parameters were as follows

·        CIBIC-Plus score at endpoint

·        Change from baseline in ADCS-ADL score at endpoint

·       The primary efficacy analysis was to be performed using the intent-to-treat population and the last-observation-carried-forward (LOCF) method of imputation (unless data from a retrieved dropout visit was available, in which case that was to be used)

·       The 2 treatment groups were to be compared using the Wilcoxon-Mann-Whitney test for independent samples; p-values and 95% confidence intervals were to be presented for treatment differences (the confidence intervals will be calculated based on normality assumptions)

·       Treatment-by-center interactions were to be evaluated in an exploratory manner, only.

11.2.6  Null And Alternative Hypotheses

The null and alternative hypotheses for each primary efficacy variable were to be tested independently. The outcome of the study was to be considered statistically significant only if both null hypotheses are rejected.

 

H0C:

Average CIBIC-Plus scores of memantine and placebo groups after 28 weeks of treatment are equal

H1C:

Average CIBIC-Plus scores of memantine and placebo groups after 28 weeks of treatment are unequal

H0A:

Average ADCS-ADL sum scores of memantine and placebo groups after 28 weeks of treatment are equal

H1A:

Average ADCS-ADL sum scores of memantine and placebo groups after 28 weeks of treatment are unequal

 

11.2.7  Additional Analyses On Primary Efficacy Parameters

Exploratory analyses were to be performed on the primary efficacy parameters, using the same statistical method as for the primary efficacy analysis, using the treated-per-protocol dataset at each timepoint (Weeks 4, 12, and 28) and the intent-to-treat dataset at Weeks 4 and 12.

11.2.8  Pooling Of Centers

The analysis plan stated that it might become necessary to pool study sites with small numbers of patients (e.g., those with £ 5 randomized patients) in order to analyze center effects (center effects on the efficacy analysis were to be analyzed on an exploratory basis only)

11.2.9  Secondary Efficacy Parameters

·       The secondary efficacy parameters were the change from baseline to each study timepoint in secondary efficacy measure scores

·       Analyses were to be performed on both the intent-to-treat and treated-per-protocol populations at each timepoint

·       The treatment groups were to be compared on the secondary efficacy parameters using the same statistical methods applied to the primary efficacy parameter

11.2.10                  Responder Analyses

·       Patients were to be classified as responders or non-responders based on their status on global, functional, and cognitive outcome measures after 28 weeks of treatment

·       Two responder definitions were to be used, based on the following criteria (all of which implied improvement or no change)

·        CIBIC-Plus score £ 4

·        Change from baseline in the modified ADCS-ADL sum score is ³ 0

·        Change from baseline in the Severe Impairment Battery score is ³ 0

·       One definition of responder satisfied all 3 criteria; the other definition of responder satisfied only the CIBIC-Plus criterion, and the ADCS-ADL or Severe Impairment Battery criterion

·       Analyses using both responder definitions were to be performed on the intent-to-treat and treated-per-protocol populations

·       Responder frequencies in the 2 treatment groups were to be compared using Fisher’s exact test

11.2.11                  Subgroup Analyses

Additional exploratory analyses of the primary efficacy parameters were to be performed for subgroups defined by age (< 75 vs > 75), sex, ApoE genotype, severity of Alzheimer’s Disease at baseline (Mini-Mental Status Examination score < 10; Mini-Mental Status Examination score ³ 10); and memantine plasma levels at endpoint

11.2.12                  Handling Of Missing Items

The methods of replacing missing items for the Severe Impairment Battery and ADCS-ADL are summarized below

11.2.12.1            Severe Impairment Battery

There are 51 separate items in this scale, with a total score ranging from 0 to 100; higher scores indicate better functioning. Single missing items were to be replaced with a “0” before calculating the total score. If more than 11 items were missing, then the total score was to be set to missing

11.2.12.2            ADCS-ADL

There are 19 separate items in this scale, with a total score ranging from 0 to 54; higher scores indicate better functioning. Single missing items were to be replaced with a “0” before calculating the total score. If more than 4 items were missing, then the total score was to be set to missing

11.2.13                  Sample Size Rationale

·       The sample size estimate was based on the CIBIC-Plus change in another memantine clinical trial, using the standard 7-point scale

·       Assumptions

·        Mean memantine-placebo difference of 0.4 points on the CIBIC-Plus at study end, with a standard deviation of 0.85 points

·        Type I error of 0.05 (2-sided)

·        Type II error of 0.05 (i.e., 95% power)

·       Based on the above assumptions, it was estimated that 107 patients would need to be randomized to each treatment group

11.3   Protocol Amendments

These have been incorporated into the outline above

11.4   Actual Analyses Performed

A supplemental statistical analysis plan is included in an appendix to the study report. It does not appear as if this plan was finalized prior to the breaking of the study blind. The key changes made to the analysis plan already described above are as follows

11.4.1  Alternative Imputation Schemes For Analysis Of CIBIC-Plus

·       In the pre-specified efficacy analysis, patients with no post-baseline CIBIC-Plus ratings were assigned a score of 4 (“unchanged”) as their endpoint rating in the LOCF dataset

·       To examine the effect of this imputation rule on the analysis results, additional endpoint analyses, using several alternate imputation schemes were conducted. These analyses were conducted after patients with missing Week 28 CIBIC-Plus scores were assigned each of the following as their endpoint assessment

·        Group mean score

·        Group median score

·        Worst case score (i.e., 7)

·        Worst group score

·       Each of the modified datasets was analyzed using a Wilcoxon-Mann-Whitney test of the difference in group means

11.4.2  Additional Analyses Of The Severe Impairment Battery

Additional analyses of the Severe Impairment Battery were conducted using the same methods as specified for the modified ADCS-ADL. These included

·       Analyses of subgroups, based on sex, age, ApoE genotype, and Alzheimer’s Disease severity at baseline

·       Analyses of treatment-by-center interactions

11.4.3  Elimination Of The Resource Utilization In Dementia Analyses From The Main Study Report

These analyses were reported separately

11.4.4  Elimination Of The Treated-Per-Protocol Analyses

Analyses using this dataset were eliminated altogether

11.4.5  Elimination Of Subgroup Efficacy Analyses Based On Plasma Levels

These analyses were eliminated altogether

11.4.6  Determination Of The Primary Reason For Discontinuation

One primary reason for treatment discontinuation was to be identified for each patient prematurely terminating the study

11.5   Efficacy Results

11.5.1  Patient Disposition

Patient disposition in this study is summarized in the following table which I have copied from the submission

 

As the study results indicate, discontinuations were more frequent in the placebo group than in the memantine group, with the majority being attributable to adverse events

11.5.2  Treatment Duration

The duration of treatment in the placebo and memantine groups is as displayed in the following 2 tables, which I have derived from tables contained in the submission. The data are based on the intent-to-treat population

 

Placebo

(n = 125)

Memantine

(n = 123)

Treatment Duration (Days)

 

 

Mean

166.1

169.58

Median

193.0

195.0

Standard Deviation

56.42

56.03

Range

3 to 218

2 to 229

 

 

 

 

 

 

Placebo

(n = 126)

Memantine

(n = 126)

Treatment Duration

 

 

1 to 30 days

9 (7%)

7 (6%)

31 to 60 days

3 (2%)

5 (4%)

61 to 90 days

7 (6%)

7 (6%)

91 to 120 days

3 (2%)

3 (2%)

121 to 150 days

5 (4%)

1 (1%)

151 to 180

20 (16%)

11 (9%)

181 to 210

71 (56%)

84 (67%)

211 to 240

7 (6%)

6 (5%)

Missing

1 (1%)

3 (2%)

The tables indicate that the majority of patients in both treatment groups received more than 180 days of study drug.

11.5.3  Demographic And Other Baseline  Characteristics

Demographic characteristics are summarized in the following table which I have copied from the submission

 

Summary statistics for baseline efficacy measures are in the following table, which I have copied from the submission

The tables above indicate that the treatment groups were broadly comparable in regard to mean age and baseline cognitive and functional status.

 

The distribution of baseline Mini-Mental Status Examination scores in the entire population enrolled in the study is in the following table:

Mini-Mental Status Examination Score

N

%

1

1

0.4

2

2

0.79

3

27

10.7

4

35

13.9

5

26

10.3

6

19

7.5

7

11

4.4

8

16

6.3

9

18

7.1

10

23

9.1

11

19

7.5

12

24

9.5

13

11

4.4

14

20

7.9

As the table above indicates, 38.4% of those enrolled in the study had a baseline Mini-Mental Status Examination score ³ 10.

11.5.4  Primary Efficacy Analysis

11.5.4.1                CIBIC-Plus

Mean CIBIC-Plus ratings at endpoint for the primary LOCF dataset and for Observed Cases (OC) are summarized in the following table, taken from the submission

 

The distribution of CIBIC-Plus ratings at endpoint for the LOCF dataset is in the following figure, which I have taken from the submission

As the table and figure above indicate, the treatment difference was clearly statistically significant only for the Observed Cases dataset; for the primary LOCF dataset, the results were borderline (p = 0.064) as regards statistical significance. For both datasets, memantine was superior to placebo.

 

Analyses of the CIBIC-Plus were also conducted using alternative imputation rules, i.e., rules that were different from those used for the LOCF analysis (these schemes are described in Section 11.4.1.). The results, which indicate a statistically significant superiority of memantine over placebo regardless of which alternative imputation scheme was used are summarized in the next table, which I have copied from the submission.

11.5.4.2                Modified ADCS-ADL

Mean change from baseline in the modified ADCS-ADL at endpoint for the primary LOCF dataset and for Observed Cases (OC) are summarized in the following table, taken from the submission

 

Changes from baseline in the ADCS-ADL (Observed Cases dataset) at each study timepoint are in the following figure taken from the submission

 

As the table and figure above indicate, there were statistically significant differences between the treatment groups on this measure for both datasets, with the memantine group being superior to the placebo group.

11.5.5  Analysis Of Secondary Efficacy Measures

11.5.5.1                Severe Impairment Battery

Mean changes from baseline in the Severe Impairment Battery at endpoint for the primary LOCF dataset and for Observed Cases (OC) are summarized in the following table, taken from the submission

Note that the exact p-value for the endpoint LOCF comparison was 0.0003

 

Changes from baseline in the Severe Impairment Battery (Observed Cases dataset) at each study timepoint are in the following figure taken from the submission

As the table and figure above indicate, there were at least nominally statistically significant differences between the treatment groups on this measure for both datasets, with the memantine group being superior to the placebo group. Although many analyses were performed in this study, apart from the primary efficacy analysis, the p-value (p = 0.0003) for the treatment group comparison on this measure on the LOCF dataset was robust enough  to remain statistically significant (i.e., p < 0.05) even after adjustment for multiple comparisons.

11.5.5.2                Other Secondary Efficacy Measures

Changes from baseline to endpoint for the other secondary efficacy parameters are in the following table which I have copied from the submission. A nominally statistically significant difference (p < 0.05) between treatment groups was seen only for Functional Assessment Staging

 

11.5.6  Additional Sponsor Analyses

The results of these analyses have been summarized by the sponsor as follows

11.5.6.1                Subscale Analyses

Analyses of individual domains/items for the CIBIC-Plus, ADCS-ADL, and Neuropsychiatry Inventory generally showed numerical trends in agreement with observations for the complete scales

11.5.6.2                Responder Analyses

Responder analyses were based on the two definitions already outlined in Section 11.2.10.

 

For the first definition, 6% of placebo patients and 11% of memantine patients were classified as responders (p = 0.170).

 

For the second definition, 10% of placebo patients and 29% of memantine patients were classified as responders (p < 0.001).

11.5.6.3                Consistency Of Treatment Effect Across Centers

In the statistical models used for analysis of the CIBIC-Plus, ADCS-ADL, and Neuropsychiatry Inventory, there was a lack of significant center effects or treatment-by-center interactions for all 3 scales (p > 0.1); the observed memantine-placebo differences at each center supported the consistency of the treatment effect across centers.

11.5.6.4                Efficacy In Subgroups

Additional exploratory analyses for subgroups defined by sex, age, severity of dementia at baseline, and presence or absence of ApoE4 allele, showed an advantage for memantine over placebo on both the LOCF and Observed Cases datasets at Week 28

11.5.6.5                Relationship Of Efficacy To Memantine Plasma Level

Plasma concentrations of memantine were determined in samples obtained from 108 memantine-treated patients at their final visit; they do not appear to have been determined in placebo-treated patients.

 

Based on their plasma levels, patients treated with memantine were grouped into 4 categories: £ 70 ng/mL; 71 – 100 ng/mL; 101 – 130 ng/mL; > 130 ng/mL.

 

The mean change from baseline in Severe Impairment Battery score in each of these categories is in the following table for both the LOCF and Observed Cases datasets.

Dataset

                        Memantine Plasma Level Category

£ 70 ng/mL

(n = 7)

71 – 100 ng/mL

(n = 28)

101 – 130 ng/mL

(n = 32)

> 130 ng/mL

(n = 41)

LOCF at endpoint

- 4.0

- 4.25

- 4.44

- 4.66

Observed Cases at Week 28

- 5.20

- 3.92

- 4.96

- 4.35

 

As the above table indicates, there was no suggestion of a correlation between memantine plasma levels and change from baseline in Severe Impairment Battery scores.

11.5.7  Agency Subgroup Analysis

Dr Tristan Massie, Agency Biometrics Reviewer of this submission, has, at my request, compared the effects of the two treatment groups on the primary efficacy parameters, after dividing those enrolled into 2 subgroups: those with a Mini-Mental Status Examination (MMSE) score ³ 10, and those with a Mini-Mental Status Examination score < 10.

 

The purpose of this additional analysis was to help determine if any effect on memantine in Alzheimer’s Disease was actually determined by patients with more severe dementia, for the following reasons

·       4 drugs have currently been approved for the treatment of mild-to-moderate dementia of the Alzheimer’s type, whereas the sponsor is currently seeking a claim for memantine in the treatment of moderate-to-severe dementia of the Alzheimer’s type. Baseline Mini-Mental Status Examination scores used to include patients in clinical trials for mild-to-moderate Alzheimer’s Disease range from 10-26; that range overlaps with the range used to select patients for MEM-MD-02

·       Patients enrolled in this study had a baseline Mini-Mental Status Examination score that ranged from 1 to 14 (with the vast majority having Mini-Mental Status Examination scores that ranged from 3 to 14, as specified by the inclusion criteria for this study). The majority of those enrolled had a Mini-Mental Status Examination score < 10.

 

The results of the analysis are summarized in the following table

 

 

 

                                         Study 9605:  ITT-LOCF MMSE Subgroup Analyses

Variable

MMSE Group

Treatment Group

n

Baseline

Mean (SD)

Mean Change From Baseline To Endpoint

Mean (SD)

p-value for treatment group comparison

Interaction

p value

Primary

 

 

 

 

 

 

 

ADL Total

<10

Placebo

73

25.5 (11.9)

-5.6 (6.5)

0.2668

0.0951

 

<10

Memantine

79

24.3 (9.0)

-4.5 (6.7)

 

³ 10

Placebo

50

30.7 (8.4)

-4.6 (6.1)

0.0095

 

³ 10

Memantine

45

31.0 (7.8)

-0.6 (6.4)

 

 

 

 

 

 

 

 

CIBIC-Plus

<10

Placebo

70

N/A

4.80 (1.06)

0.5364

 

 

<10

Memantine

75

N/A

4.68 (1.10)

 

³ 10

Placebo

48

N/A

4.75 (1.14)

0.0231

 

³ 10

Memantine

43

N/A

4.23 (1.09)

 

 

 

 

 

 

 

 

Secondary

 

 

 

 

 

 

 

SIB Total

<10

Placebo

73

58.0 (19.4)

-11.8 (14.0)

0.0091

0.8136

 

<10

Memantine

79

55.0 (20.4)

-5.8 (12.6)

 

³ 10

Placebo

50

83.7 (8.8)

-7.6 (12.5)

0.0087

 

³ 10

Memantine

45

84.8 (11.3)

-0.8 (7.9)

As the table above indicates, differences between treatment groups (effect sizes) appeared to be greater for those with a baseline Mini-Mental Status Examination ³ 10, for both primary measures (and to a lesser extent for the Severe Impairment Battery)

11.6   Sponsor’s Conclusions Regarding Efficacy

  • A statistically significant superiority of memantine over placebo was observed for the ADCS-ADL and Severe Impairment Battery on the LOCF analysis at endpoint, and for the Observed Cases analysis at the same timepoint
  • A marginally significant superiority of memantine over placebo was observed for the CIBIC-Plus on the LOCF analysis at endpoint. However, a clearly statistically significant advantage was observed for the Observed Cases analysis at Week 28. The robustness of the analysis of the CIBIC-Plus was further supported by analyses using alternative imputation rules

11.7   Agency Statistical Reviewer’s Comments

Final comments from the Agency statistical reviewer are pending.

11.8   Reviewer’s Comments

·       This study was intended to evaluate the efficacy of memantine compared with placebo in moderate-to-severe Alzheimer’s Disease. The study had 2 primary efficacy measures, the CIBIC-Plus (a global measure) and the modified ADCS-ADL (a measure of activities of daily living). The prospectively-finalized analysis plan indicated, that for the study to be declared positive, a statistically significant difference (p < 0.05) between memantine and placebo needed to be seen on both primary efficacy measures, using the prospectively-specified dataset and analytical method.

·       The protocol-specified primary analysis, on the LOCF dataset, provided a borderline level of statistical significance for the CIBIC-Plus (p = 0.064) and clear statistical significance for the modified ADCS-ADL (0.022), when the 2 treatment groups were compared. More clearly statistically significant results were seen for both parameters when the Observed Cases (at Week 28) dataset was analyzed.

·       Thus far, the regulatory standard for determining the efficacy of drugs intended for the treatment of Alzheimer’s Disease/dementia of the Alzheimer’s type has been the demonstration of a statistically significant (p < 0.05) advantage for the drug in comparison with placebo on 2 types of primary efficacy measure: a cognitive measure, since cognitive dysfunction is the core manifestation of dementia; and a global or functional measure, so as to confirm that any effect on the cognitive measure is clinically meaningful

·       This study lacks a cognitive primary efficacy measure; in designing this protocol, the original sponsor took the view that demonstrating efficacy on global and functional measures was more practical and meaningful than demonstrating efficacy on a cognitive measure, in a population with severely impaired cognition

·       The study does however have a secondary efficacy measure (one of seven), the Severe Impairment Battery, that is specifically intended to measure change in cognition in patients with severe dementia. An at least nominally statistically significant difference (p = 0.0003) between memantine and placebo was seen on this measure for the LOCF dataset at study endpoint; this p-value appeared robust enough to remain statistically significant (p < 0.05) when adjusted for multiple comparisons.

·       Thus this study could be considered to have shown evidence of a statistically significant superiority for memantine over placebo on both a cognitive and a global primary efficacy measure, and to be consistent with the regulatory standard for determining the efficacy of drugs in Alzheimer’s Disease/dementia of the Alzheimer’s type.

·       The following are also noteworthy, however

·        In both the memantine and placebo groups there was a mean deterioration in cognitive function over the 28-week course of the study

·        The effect size on the Severe Impairment Battery remained relatively small (5.91 point mean difference between treatment groups on the Severe Impairment Battery for the LOCF dataset at study end)

·        Based on the response patterns seen on the CIBIC-Plus, only a small minority of patients treated with memantine showed even a minimal or moderate improvement, with no patients showing a marked improvement, and the most common response being “no change”

12    Study MRZ 9403

This study was conducted at 7 centers in Latvia.

12.1   Title

Efficacy And Tolerability Of Akatinol Memantine In Care-Dependent Patients With Moderate To Severe Primary Dementia

12.2   Objective

To evaluate the clinical efficacy and tolerability of memantine in care-dependent patients with moderate-to-severe dementia

12.3   Design

Randomized, double-blind, placebo-controlled, parallel-arm study

12.4   Duration

12 weeks of double-blind treatment

12.5   Dosage

The dosing regime for this study was as follows

Study Days

Dosage

1 to 7

Memantine 5 mg or matching placebo once daily in the morning

8 to 84

Memantine 10 mg or matching placebo once daily in the morning

 

12.6   Sample Size

150 patients were to be enrolled in the study and randomized equally to the two treatment groups

12.7   Main Inclusion Criteria

·       Male or female

·       Age: 60 to 80 years

·       Resident in a nursing home

·       Education up to at least the elementary school level

·       Moderate-to-severe dementia based on the DSM-III-R and the following criteria

·        Global Deterioration Scale: 5 to 7 points

·        Clinical Global Impression of Severity score of 5 to 7

·        Mini-Mental Status Examination score < 10

 

Note that the original study protocol states that patients targeted for enrollment in this study were to include those with Alzheimer’s Disease, vascular dementia, and mixed dementia (combining Alzheimer’s Disease with vascular dementia); criteria for making these diagnoses at study entry are not specified. The original study protocol further states the following: “As patients with both (sic) types of dementia are to be included in the trial, the results of a CT examination and a Hachinski Ischemic Scale test done at the beginning of the trial will NOT (emphasis mine) be utilized to differentiate between primary degenerative dementia and vascular dementia.”

·       Duration of dementia or symptoms > 12 months

·       No “clinically relevant pathological changes” in the following laboratory data (taking into consideration age-related alterations): CBC, electrolytes, BUN, serum creatinine, GGT, ALT, total protein, and urinalysis

·       No clinically relevant reductions in serum vitamin B12 or in thyroid functions

·       No central nervous system active drugs taken within 14 days before the trial

·       Informed consent

12.8   Main Exclusion Criteria

·       Severe hypothyroidism and other relevant endocrine diseases

·       Unstable diabetes mellitus

·       Severe chronic or terminal diseases

·       Cardiac failure (NYHA Class III or IV)

·       Severe fixed hypertension (WHO Class III) or labile hypertension while under treatment

·       Myocardial infarction, endocarditis, or myocarditis during the last 3 months

·       Severe arrhythmias requiring treatment

·       Severe orthostatic “dysregulation”

·       Severe chronic obstructive pulmonary disease

·       Chronic liver disease (transaminases > 2 x upper limit of normal); hepatic encephalopathy

·       Severe renal disease or dysfunction (serum creatinine > 2 mg/dL)

·       Brain tumor

·       Schizophrenia

·       Major depression (Hamilton Depression Scale [21-item version] score > 18)

·       “Oligophrenia”

·       Epilepsy

·       Parkinson’s Disease

·       Secondary dementia

·       Alcoholism, drug addiction

·       Participation in a clinical trial within the preceding 30 days

·       Blood loss of > 500 mL within the preceding 2 months

·       The following concomitant medications

 

·        Medications with could interact with the study drug or influence the results of efficacy testing (these were to be withdrawn 14 days before the start of the trial, and were not to be administered during the trial)

·        Anticonvulsants

·        Monoamine oxidase inhibitors, neuroleptics, tricyclic antidepressants

·        Nootropics or agents stated to promote cerebral circulation

·        Hypnotics, except for chloral hydrate or benzodiazepines with short half-lives

12.9   Concomitant Medications

12.9.1  Prohibited Medications

The following concomitant medications are prohibited (as already noted)

 

·       Medications with could interact with the study drug or influence the results of efficacy testing (these were to be withdrawn 14 days before the start of the trial, and were not to be administered during the trial)

·       Anticonvulsants

·       Monoamine oxidase inhibitors, neuroleptics, tricyclic antidepressants

·       Nootropics or agents stated to promote cerebral circulation

·       Hypnotics, except for chloral hydrate or benzodiazepines with short half-lives

12.9.2  Permitted Medications

Long-term treatment with drugs such as cardiac glycosides, antihypertensives and oral anti-diabetic agents is permitted as long as dosage is kept constant before and during the clinical trial phase

12.10    Schedule

Study visits were to be at screening/baseline (no clear distinction is made in the protocol between the screening and baseline visits) and Days 7, 28, 56, and 84.

 

The study schedule is summarized in the following table

 

Day

0

7

28

56

84

History

X

 

 

 

 

Physical examination

X

 

 

 

X

Risk factor data

X

 

 

 

X

Neurological examination

X

 

 

 

 

Memantine plasma concentration

X

 

X

 

X

Safety laboratory tests

X

 

X

 

X

Hachinski

X

 

 

 

 

CT scan of brain (optional)

X

 

 

 

 

DSM-III-R

X

 

 

 

 

GDS

X

 

 

 

 

MMSE

X

 

 

 

 

CGI-C

X

 

X

 

X

CGI-S

X

 

 

 

X

CGI Benefit/Risk Index

X

 

 

 

X

G2

X

 

 

 

X

G2-C

 

X

X

X

X

BGP

X

X

X

X

X

IADLPT

X

 

 

 

X

Medication compliance

 

X

X

X

X

Medication dispensation

X

X

X

X

 

Adverse events

 

X

X

X

X

Blood pressure, heart rate

X

X

X

X

X

 

12.11    Outcome Measures (Per-Protocol)

12.11.1                  Primary Efficacy Measures

BGP Care Dependency Subscale

CGI-C (dichotomized): responder rate

(note that a responder is not clearly defined in the protocol)

12.11.2                  Secondary Efficacy Measures

IADLPT (timing and quality)

G2 (single item scores and total score)

G2-C (single item scores and total score)

12.11.3                  Safety Measures

(The analysis of these measures will be not be further addressed here, as this is an efficacy review)

Adverse events

Safety laboratory tests (hematology, clinical chemistry and urinalysis)

12.11.4                  Pharmacokinetic Measures

Plasma levels of memantine

12.12    Analysis Plan (Per-Protocol)

12.12.1                  General Considerations

·       A Type I error of 0.025 (2-sided) was to be used

·       Results were to be presented using descriptive statistics

12.12.2                  Demographic And Baseline Characteristics

No details are supplied

12.12.3                  Study Hypotheses

H0(1): There are no differences at the end of treatment between memantine and placebo with regard to the responder rate on the basis of the dichotomized CGI-C

H1(1): There are differences at the end of treatment between memantine and placebo with regard to the responder rate on the basis of the dichotomized CGI-C

 

H0(2): There are no differences at the end of treatment between memantine and placebo with regard to the BGP Care Dependency Subscale change from baseline score

H1(2): There are differences at the end of treatment between memantine and placebo with regard to the BGP Care Dependency Subscale change from baseline score

12.12.4                  Primary Efficacy Parameters

·       The primary efficacy parameters were to be as follows

·        Change from baseline to endpoint in the BGP Care Dependency Subscale

·        CGI-C (dichotomized) responder rate at study endpoint

·       The population for the primary efficacy was “intent-to-treat,” defined as all those who received study medication and had Day 28 measurements while taking study medication

·       Differences between the 2 treatment groups on the BGP Care Dependency Subscale were to be analyzed using Wilcoxon-Mann-Whitney U tests

·       Differences between treatment groups on the CGI-C were to be analyzed using Fisher’s exact test

·       Missing data were to be replaced using “worst ranks”

12.12.5                  Secondary Efficacy Parameters And Other Analyses

·       Secondary efficacy variables, and the residual results of the CGI-C and BGP, were to be checked for medication and time effects, as well as for interactions using suitable non-parametric methods

·       Subgroup analyses, based on age and severity of disease, were to be done using the relevant frequency distributions.

·       If the sample was big enough, descriptive analyses for center effects were also intended

12.12.6                  Sample Size Calculation

12.12.6.1            For CGI-C

·       Assumptions

·        Type I error: 0.025

·        Power: 90%

·        30% difference in responder rate on the CGI-C between the treatment and placebo groups; responder rate 30% in placebo group (on dichotomized scale).

·       Based on the above assumptions a sample size of 68 patients per treatment group was estimated

12.12.6.2            For BGP Care Dependency Subscale

·       Assumptions

·        Type I error: 0.025

·        Power: 90%

·        7.8 point difference in the change from baseline on the BGP care dependency subscale.

·       Based on the above assumptions a sample size of 23 patients per treatment group was estimated

12.12.6.3            Overall

Based on the above sample size calculation, a total enrollment of 136 patients was estimated

12.12.7                  Interim Analysis

None planned.

12.13    Protocol Amendments

The following key amendments were made to the protocol prior to the study blind being broken

·       Introduction of 6 additional study centers

·       An increase in total number of patients randomized to 168

12.14    Post-Hoc Analysis Plan (Forest Laboratories)

The study was completed by Merz in 1995, and the results published in 1999 as follows.

 

Winblad B, Poritis N. Memantine in severe dementia: results of the 9M-Best Study (Benefit and efficacy in severely demented patients during treatment with memantine). Int J Geriatr Psychiatry 1999;14:135-46

 

A new analysis plan was finalized by Forest Laboratories on May 24, 2002. This analysis plan is further described below. The analysis described in the study report is based on this post-hoc analysis plan

12.14.1                  Objectives

12.14.1.1            Primary

To evaluate the efficacy, safety, and tolerability of memantine as compared with placebo in patients with moderate-to-severe dementia of Alzheimer’s and vascular type

12.14.1.2            Secondary

·       To further compare the efficacy of memantine relative to placebo using several secondary efficacy parameters

·       To assess the safety and tolerability of memantine (this appears to be a primary as well as secondary objective)

12.14.2                  Efficacy Outcome Measures

12.14.2.1.1            Primary

·       CGI-C (7-point scale);

Data for the dichotomized CGI-C responder analysis were also to be presented, but the analysis of the 7-point scale was to be primary

·       BGP Care Dependency Subscale

·       BGP Cognitive Subscale

12.14.2.1.2            Secondary

·       BGP Total Score and all other BGP sub-scales

·       CGI Efficacy Index and CGI Risk Index

·       CGI-S

·       G2; G2-C

·       IADLPT (timing and quality)

 

12.14.3                  Study Populations

The sponsor has defined the following patient populations for purposes of analysis

12.14.3.1            Randomized Population

This population was to consist of all patients randomized into the study

12.14.3.2            Safety Population

This population was to consist of all randomized patients who received at least one dose of double-blind study medication

12.14.3.3            Intent-To-Treat Population

This population was to consist of all those in the safety population who completed at least one post-baseline efficacy evaluation of the CGI-C or BGP. Missing data were to be imputed when an analysis was performed on this population

12.14.4                  Patient Disposition And Study Completion

·       The number of patients in each study population (i.e., randomized, safety, intent-to-treat) were to be summarized by treatment group and center

·       The number of patients with Alzheimer’s Disease and the number of patients with vascular dementia in each study population were to be presented by treatment group and center

·       The number and percentage of the total population, as well in each dementia subtype population (i.e., Alzheimer’s Disease and vascular dementia) completing and discontinuing during the double-blind treatment period were to be presented by treatment group. Reasons for discontinuation were to be presented by treatment group

12.14.5                  Demographic And Other Baseline Characteristics

·       Demographic parameters and other baseline characteristics were to be summarized by treatment group

·       The treatment groups were to be compared as follows

·     Continuous variables were to be analyzed using a 2-way ANOVA model with treatment and study center as the factors

·     Categorical variables were to be analyzed using a Cochran-Mantel-Haenszel test controlling for study center

12.14.6                  Efficacy Analyses

12.14.6.1            General

·       All efficacy analyses were to be based on the intent-to-treat population

·        Primary analyses were to be performed using the LOCF approach: the change score from baseline to Week 24 will be used

·        Supportive analyses were to use the Observed Cases and Worst Case approaches

·        Descriptive statistics were to be calculated for each visit using both approaches

·       All statistical tests were to be 2-sided and a p-value of < 0.05 was to be considered statistically significant for main effects, and 10% for interaction terms

12.14.6.2            Primary Efficacy Parameters

·       The two primary efficacy parameters were to be the following

·        CGI-C score at endpoint (based on original 7-point scale) [data for the responder analysis of the dichotomized CGI-C scale was also to be presented]

·        Change from baseline to endpoint in BGP Care Dependency Subscale

·       Another “key” parameter of efficacy (also considered a primary efficacy parameter) was to be the BGP Cognitive Subscale

·       The primary efficacy analysis was to use the intent-to-treat population with the last-observation-carried-forward (LOCF) method of imputing missing data.

·       The original 7-point CGI-C scale was to be analyzed using the stratified (by center) Wilcoxon rank-sum test. The dichotomized CGI-C was to be analyzed using Fisher’s exact test and the stratified (by center) Wilcoxon rank-sum test.

·       The BGP Care Dependency Subscale and the BGP Cognitive Subscale were to be analyzed using the stratified (by center) Wilcoxon rank-sum test.

·       Since treatment superiority needed to be shown on all 3 primary efficacy parameters (p < 0.05), no multiplicity adjustment was felt to be necessary.

12.14.6.3            Sub-Population Analyses

Those with a modified Hachinski Ischemic Scale score of £ 4 were identified as having dementia of the Alzheimer’s type. The primary efficacy analyses on the total population were to be repeated on this subset.

12.14.6.4            Secondary Efficacy Parameters

·       Analyses of the secondary efficacy parameters were to use the same statistical methods that were used for the primary efficacy analyses

·       Analyses were to use the intent-to-treat-LOCF population, with supportive analyses using the Observed Cases and Worst Case datasets

12.14.6.5            Additional Analyses

·       By-center descriptive analyses for the 3 key efficacy parameters were to be provided to assess center consistency

·       Descriptive analyses of three key efficacy variables were to be provided based on gender, age group (< 75, ³ 75), and baseline BGP Care Dependency Subscale score (< 20 or ³ 20)

·       A correlation analysis was to be conducted to assess the extent to which changes in the BGP Care Dependency Subscale score were attributable to changes in the BGP Cognitive Subscale score.

12.14.6.6            Handling Of Missing Data

·       Missing values for efficacy variables were to be imputed using the following methods

·        Last-observation-carried-forward (LOCF): The last observed value prior to the missing value was to be used

·        Worst case: Imputation was to be based on the worst rank  for each efficacy parameter, as depicted in the following table

Efficacy Parameter

Worst Rank

CGI-C (7-point scale)

7

CGI-C (Dichotomized)

Non-responder

BGP Care Dependency Subscale

46

BGP Cognitive Subscale

10

BGP Total

70

BGP Aggressiveness

10

BGP Depression

6

BGP Mental Disability

8

BGP Inactivity

12

G2 Total

102

G2 Item

6 per item

G2-C Total

112

G2-C Item

7 per item

CGI-S

7

 

12.14.7                  Exposure And Dosing Compliance

·       The safety population will be used for both exposure and study medication compliance.

·       Double-blind medication exposure will be calculated as the difference between the date when double-blind medication was first taken, and the date when the last dose was taken (i.e., total days dosed) plus 1.

·       Study medication compliance is calculated as the total number of tablets taken by a patient during the patient’s participation in the double-blind medication phase divided by the number of tablets expected to be taken during that period, multiplied by 100. Overall, compliance rates £ 75% of double-blind medication are considered compliant.

·       Descriptive statistics for study medication compliance rate and frequency distribution for the number of compliant patients will be presented by treatment group for the double-blind study period.

12.14.8                  Sample Size Estimate

12.14.8.1            For CGI-C

·       Assumptions

·        Type I error: 0.025

·        Power: 90%

·        30% difference in responder rate on the CGI-C between the treatment and placebo groups; responder rate 30% in placebo group (on dichotomized scale).

·       Based on the above assumptions a sample size of 68 patients per treatment group was estimated

12.14.8.2            For BGP Care Dependency Subscale

·       Assumptions

·        Type I error: 0.025

·        Power: 90%

·        7.8 point difference in the change from baseline on the BGP care dependency subscale.

·       Based on the above assumptions a sample size of 23 patients per treatment group was estimated

12.14.8.3            Overall

Based on the above sample size calculation, a total of 136 patients completing the study was estimated. Assuming a 10% dropout rate, 150 patients per treatment group were estimated to be needed

12.15    Key Changes Contained In Post-Hoc Analysis Plan

The following were the key changes contained in the post-hoc analysis plan, as drawn up in 2002, as compared with the original protocol and analysis plan that was drawn up prior to the study blind being broken

 

·       The primary efficacy analysis was to use the LOCF approach for imputing missing data, rather than the Worst Case approach

·       The 7-point CGI-C scale was to be used for the primary efficacy analysis, rather than the responder analysis of the dichotomized scale

·       The BGP Cognitive Subscale, a subset of the BGP Care Dependency Subscale, was to be included as a “key” (i.e., primary) efficacy measure.

·       The primary efficacy analysis was also to be performed on the dementia of the Alzheimer’s type subset as defined by a modified Hachinski Ischemic Scale score £ 4

·       The per-protocol dataset was eliminated from the efficacy analysis

·       The method for imputing the worst possible change from baseline on the BGP Care Dependency Subscale was altered as follows

·        Scores on this scale range from 0 (best) to 46 (worst)

·        In the original statistical analysis plan, when the post-baseline measurement was missing, a change score of 46 was imputed, implying that the baseline value was considered to be zero; i.e., the true baseline value was not used

·        In the post-hoc analysis plan, the missing value was set to 46, but the observed baseline value was not replaced

·        The same method was used for imputing all data related to the BGP

12.16    Efficacy Results

12.16.1                  Patient Disposition

166 patients were randomized; their disposition, according to dementia subgroup was as follows (as noted earlier, those with a modified Hachinski Ischemic Scale £ 4 were considered to have dementia of the Alzheimer’s type, where those with a score > 4 were considered to have vascular dementia). Randomization was NOT stratified by dementia subgroup

 

 

                                              PLACEBO

                                      MEMANTINE

DAT

n

VAD

n

Total

n

DAT

n

VADn

Total

n

Randomized

38

46

84

41

41

82

Completed

37

43

80

39

39

78

Discontinued

1

3

4

2

2

4

DAT: Dementia of the Alzheimer’s type; VAD: vascular dementia 

 

All discontinuations were due to adverse events.

12.16.2                  Protocol Deviations

2 patients in each treatment group entered the study despite not satisfying eligibility criteria based on age, laboratory abnormalities or age; these included one patient in the placebo group with cirrhosis.

12.16.3                  Demographic And Other Baseline Characteristics

These are summarized in the following table

Variable

Placebo

(n = 84)

Memantine

(n = 82)

Males (%)

44.0

40.2

Mean Age (years)

71.9

71.2

Mean Weight (kg)

67.4

67.9

Mean MMSE Score

6.1

6.5

Mean GDS Score

6.0

6.0

Mean CGI-S Score

5.7

5.5

Mean Hachinski Ischemic Scale Score

5.7

5.2

Mean Hamilton Depression Scale Score

8.9

8.5

Mean BGP Care Dependency Subscale Score

21.8

21.3

Mean BGP Cognitive Subscale Score

5.4

5.5

 

 

As the table above indicates, the treatment groups were largely comparable at baseline.

 

Note, that the mean modified Hachinski Ischemic Scale score at baseline was above 4 in both treatment groups; further data regarding the distribution of this measure among the 2 treatment groups was as follows.

Variable

Placebo

(n=84)

Memantine

(n = 82)

Mean Hachinski Ischemic Scale Score

5.7

5.2

Median Hachinski Ischemic Scale Score

5.0

4.5

Standard Deviation

3.2

2.9

Range

1 to 12

1 to 12

 

12.16.4                  Brain Imaging At Study Entry

Only a total of 86 patients enrolled in this study had brain imaging at study entry. Their CT scan reports (translated into English) were provided to this Division on request.

 

I have read these reports in detail and have attempted to find patients whose radiological findings suggested a possible cause for dementia other than Alzheimer’s Disease, vascular dementia, or mixed dementia.

 

Note that patients were grouped post-hoc into 2 categories based on their modified Hachinski Ischemic Scale score rather than based on their CT scan reports.

 

All CT scans were done without contrast.

 

CT scan reports which suggested a possible etiology for dementia separate from, or in addition to, a primary degenerative dementia and/or vascular dementia are as follows

 

Patient #; Initials

CT scan report

011

Quite remarkably enlarged ventricular system.

Osteoplasty of the right temporal-parietal bone after craniotomy; metallic blood vessel clips on the dura mater.

There is large area of encephalomalacia in the left temporal lobe – sequelae of previous cranial trauma

Heavily calcified syphon parts of both carotid arteries

 

Conclusion; Atrophic changes in the brain due to cranial trauma and atherosclerosis

 

064

The 4th and 3rd ventricles are localized in the midline. The enlarged lateral ventricles are symmetrically localized. The anterior horn of the left lateral ventricle is retracted anteriorly

The subarachnoid spaces are enlarged

There is a liquor density space (approx 5 x 5 cm in the axial plane) in the left parietal lobe, localized against the medial part of the lateral ventricle

The bone fragment in the place of surgical operation is mildly pressed out

 

Conclusion: Moderate to marked atrophic changes in the brain. Porencephalic cavity in the left parietal lobe communicates with the lateral ventricle.

 

124

The 4th and 3rd ventricles are positioned in the midline and enlarged, more so the third ventricle. The lateral ventricles are symmetrically localized.

There is hypodense (liquor isodense) area (approx  2 x  2 cm in size) in the right parietal lobe towards the occipital horn and communicates with it.

There is hyperdense area in the region of calvarium

 

Conclusion: Atrophic changes of the brain. The cystic lesion towards the right occipital horn with greater possibility could be sequelae of head trauma

 

 

12.16.5                  Extent Of Exposure And Compliance

The mean treatment duration was 82.3 days (standard deviation 9.1 days) and 81.9 days (standard deviation 9.6 days) in the placebo and memantine treatment groups, respectively.

 

All patients in both treatment groups were considered compliant, based on pre-specified criteria

12.16.6                  Primary Efficacy Analysis

The analysis of the original two primary efficacy parameters, as well as an additional efficacy measure designated post-hoc as key and primary, are described in this section

12.16.6.1            CGI-C

Mean CGI-C scores, on the 7-point scale, at Week 12 in each treatment group are in the following table, which depicts the results for each dataset.

 

Dataset

     Placebo

     Memantine

p-value

Memantine vs placebo

N

Mean (± SEM)

N

Mean (± SEM)

LOCF

84

3.5