Alzheimer's disease - donepezil, galantamine, rivastigmine and memantine

173 Pages·2010·2.15 MB·English

The effectiveness and cost-effectiveness
of donepezil, galantamine, rivastigmine
and memantine for the treatment of
Alzheimer’s disease (review of TA111): a
systematic review and economic model
Produced by: Peninsula Technology Assessment Group (PenTAG), Peninsula Medical
School, University of Exeter
APPENDICES
APPENDIX 1: OUTCOME MEASURES............................................................................................................................. 2
APPENDIX 2: LITERATURE SEARCH STRATEGIES .................................................................................................... 11
APPENDIX 3: DATA EXTRACTION FORMS .................................................................................................................. 19
APPENDIX 4: FUNNEL PLOTS FROM THE SYNTHESIS WITH EXISTING EVIDENCE ................................................ 73
APPENDIX 5: COMBINED DOSE AND DOSE-SPECIFIC META-ANALYSES ............................................................... 75
APPENDIX 6: DATA SETS USED IN META-ANALYSIS OF POOLED MULTIPLE OUTCOME MEASURES ................ 90
APPENDIX 7: META-REGRESSION FIGURES .............................................................................................................. 94
APPENDIX 8: WINBUGS CODE FOR MIXED TREATMENT COMPARISONS ............................................................. 103
APPENDIX 9: MIXED TREATMENT COMPARISONS PERFORMED IN SPECIFIED MEASUREMENT
POPULATIONS ...................................................................................................................................... 104
APPENDIX 10: STUDIES INCLUDED BY INDUSTRY BUT EXCLUDED FROM THE PENTAG CLINICAL
EFFECTIVENESS SYSTEMATIC REVIEW ............................................................................................ 116
APPENDIX 11: ONGOING TRIALS ................................................................................................................................. 125
APPENDIX 12: PRISMA STATEMENT CHECKLIST ...................................................................................................... 132
APPENDIX 13: SUMMARY TABLES OF RESULTS FROM THE INSTITUTE OF QUALITY AND EFFICIENCY IN
HEALTH CARE. ..................................................................................................................................... 134
APPENDIX 14: MEMANTINE  ACHEI V. PLACEBO  ACHEI ...................................................................................... 136
APPENDIX 15: UPDATE ON EVIDENCE ABOUT THE CARE COST OF ALZHEIMER’S DISEASE IN THE UK ........... 143
APPENDIX 16: CONSIDERATION OF A TWO-DIMENSIONAL MARKOV MODEL FOR ALZHEIMER’S DISEASE ...... 150
APPENDIX 17: PREVIOUS CRITICISMS OF THE SHTAC ALZHEIMER’S DISEASE MODEL ...................................... 155
APPENDIX 18: PUBLISHED UTILITY VALUES FOR ALZHEIMER’S DISEASE ............................................................ 158
APPENDIX 19: FIGURES FROM THE STATISTICAL ANALYSIS OF IPD FROM WOLSTENHOLME AND
COLLEAGUES ....................................................................................................................................... 160
APPENDIX 20: GRAPHICAL PRESENTATION OF DISTRIBUTIONS FOR PSA ........................................................... 163
APPENDIX 21: TORNADO PLOTS FOR ACHEI VERSUS BEST SUPPORTIVE CARE ................................................ 166
REFERENCES TO APPENDICES ......................................................................................................................................... 169
AChEIs & memantine for Alzheimer's Appendices

Appendix 1: Outcome measures
These Tables of outcome measures have been copied from the previous TAR, TA 111,
Appendix 6.1
Global outcome measures
Type Construct measure and scoring Critical appraisal
Clinical Dementia Cognitive impairment in memory, Provides physicians with a global
Rating orientation, judgement/problem- rating that encompasses a broad
(CDR) and Clinical solving, community affairs, range of patient characteristics and
Dementia Rating Sum home/hobbies, and personal care can be used by neurologists,
of Boxes 0=none, 0.5=questionable, 1=mild, psychiatrists, and psychologists and
(CDR-SB) 2=moderate, 3=severe focuses on cognition, not on items that
may be related to other medical,
CDR-SB is a modified form which emotional or social conditions.
sums the ratings in the six Good inter-rater reliability and fair to
performance categories to give a good concurrent validity. Although no
global dementia ranking. work has been done on test-retest
reliability, nothing so far suggests that
researchers should avoid this scale
when trying to stage AD. The CDR
can be used as an eligibility criterion
for trial participation or as an outcome
measure.
Global Deterioration Progressive stages of cognitive Most frequently used but ratings can
Scale impairment misstate a patient‟s severity.
(GDS) 1 (no cognitive decline)-7 (very severe Problems might arise when the GDS
cognitive decline) is used as an inclusion criterion for
participation in an RCT. The ability to
enrol desired patients could be
threatened if the GDS misidentifies the
stages of dementia.
The GDS should not be used to stage
dementia in Alzheimer‟s Disease drug
trials.
Clinical Global Overall improvement in patient health Fair to good test-retest and inter-rater
Impression of Change status assessed by clinician (-with reliability and concurrent validity.
scale (CGIC) and the caregiver) Results may arise from fact that
global improvement 1 (very much improved) - 7 (very groups providing global assessments
index with much worse) do not base their ratings on the same
interviewing of domains. Physicians take clinical
patients Clinician A number of different variations are psychopathology as the basis of
Interview-Based available determining global improvement,
Impression of Change nurses believe the amount of work
(CIBIC) and with Scale is nonparametric and of a non- needed to care for patients was
caregiver input interval nature. important. This instrument also
Confidential material removed PenTAG 2010
- 2 -
AChEIs & memantine for Alzheimer's Appendices

Type Construct measure and scoring Critical appraisal
(CIBIC-M or –Plus) includes a caregiver opinion, results
may differ depending on whether the
rater first interviews the patient or
caregiver. The number of different
variations may have reduced the
validity.
Gottfries-Bråne-Steen Motor function, intellectual function, Psychometric properties range from
(GBS) emotional function and symptoms fair to good. Scale is useful mean of
common to demented patients. quantifying dementia in drug trials.
0 (normal function or absence of GBS should not be used as a
symptoms) to 6 (maximal disturbance diagnostic tool.
or presence of symptoms)
Mental Function A modification of the GBS prepared by Unable to source data on reliability
Impairment Scale the study authors for a previous study. and validity.
(MENFIS) Scores range from 0 to 78, with a
higher score indicating a greater
degree of deficit.
Patient Global 7 point Likert scale ranges from 1 Unable to source data on reliability
Assessment (PGA) (very much improved) to 4 (no and validity.
change) to 7 (very much worse)
Cognitive outcome measurement scales
Type Construct measure and scoring Critical appraisal
Alzheimer‟s Disease Orientation, memory, language and Limited in its ability to detect change at
Assessment Scale- praxis one end or the other of the severity
cognitive 0-70, with higher scores indicating continuum. For many subtests,
(ADAS-cog) greater impairment detection of improvement appears
only possible for a restricted range of
severity levels.
Limitations should be considered
when used as a drug efficacy
measure. The rate of decline of AD
using ADAS-cog suggests that the
decline is non-linear and not a
constant but is dependent on the
stage of the disease. Content and
ecological validity are lacking.
Benton Visual Assesses visual perception, visual The interscorer agreement for total
Retention Test memory, and visuoconstructive error score is high and for major
(BVRT) abilities. The test has three alternate categories of errors reliability is
forms, each consisting of ten designs. moderate to high. A correlation of 0.42
In addition, there are four possible was found between the Benton and
modes of administration. Scoring is the Digit Span WAIS subtest. This low
based on an assessment of the correlation indicates discriminate
number and types of errors made validity since the Benton was created
compared with the expected scores to supplement the Digit Span test.
found in the norm tables. The wider
Confidential material removed PenTAG 2010
- 3 -
AChEIs & memantine for Alzheimer's Appendices

Type Construct measure and scoring Critical appraisal
the discrepancy in favour of the Educational level may influence a
expected score, the more probable it participant‟s score on the test.
is that the participant has suffered Participants with higher educational
neurological impairment. levels tend to use a more exhaustive
exploration strategy during the
recognition phase of the test, allowing
them to perform better than
participants with lower educational
levels. The executive working memory
component is more efficient in
participants with higher educational
levels.
Computerised A computerized version of the Memory The MAC-Q questionnaire
Memory Battery Assessment Clinical Battery (MAC) demonstrates internal consistency and
(CMBT) designed to simulate critical cognitive test-retest reliability.
tasks: Name-Face Association
(delayed recall and total
acquisition);First and Last Names
(total acquisition), Facial Recognition
(first miss and total correct);
Telephone Number Recall (7-digit and
10-digit number correct); House and
Object Placement Task (total
acquisition and first trial)
Clinical Global This rating instrument expresses the This is a sub-test of the CGI, it is easy
Impression-item 2 global change in observable cognitive and quick to administer and is widely
(CGI- 2) functioning directly on a transitional used in clinical and trial settings.
scale ranging from 1 (very much
improved) to 7 (very much
deteriorated) as rated by a clinician.
Digit symbol Participants fill in a grid of 100 blank Performance on this test is affected by
substitution subtest squares, each paired with a randomly many different components, so the
(DSST) of the assigned number from 1 to 9, using a test lacks specificity. Participants with
Wechsler Adult key that pairs each number with a impaired vision or visuomotor
Intelligence Scale- different symbol. The score is the coordination, pronounced motor
Revised number of correct answers after 90 slowing or low education levels are at
seconds. a disadvantage.
Fuld object-memory Ten item assessment with ten Unable to source data on reliability
evaluation (FOME) common objects in a bag are and validity.
presented "to determine whether the
patient can identify objects by touch"
(stereognosis). The test was
developed while testing large samples
of aged adults, nursing home
residents and community active
people, for whom norms are provided.
Mini-Mental State 11 questions on orientation, memory, Good reliability and validity for its
Examination (MMSE) concentration, language and praxis. original purpose of screening for
Confidential material removed PenTAG 2010
- 4 -
AChEIs & memantine for Alzheimer's Appendices

Type Construct measure and scoring Critical appraisal
Scale ranges from 0-30. Higher score dementia, short screening scales are
indicates less impairment. There is no not designed to measure more subtle
range of scores that can be rigidly and aspects of cognition. Short scales
universally applied to indicate such as the MMSE may indicate little
dementia severity i.e. as a marker of or no change over time in subjects
mild, moderate and severe dementia. who would otherwise be shown to
In clinical trials often a score of 21-26 have declined substantially if another
is associated with mild AD, moderate scale had been used to measure
AD is associated with an MMSE of 10 change in status. Not an ideal
to 20 and severe AD is usually outcome measure for AD drug trials,
associated with an MMSE of less than especially if the expected benefits are
10. This may be less suitable within not large. It has dependence on intact
routine daily practice. language ability and there are no
available validated versions in
languages suitable for use with ethnic
minorities. It cannot be used
effectively in people with low IQs or
learning disabilities.
Severe Impairment A measure of cognition that was The SIB has been shown to be
Battery (SIB) developed to assess a range of psychometrically reliable and clinical
cognitive functioning in individuals who norms are available. No further details
are too impaired to complete standard of reliability and validity have been
neuropsychological tests and takes sourced.
into account specific behavioural and
cognitive deficits associated with
severe dementia. It is composed of 40
simple one-step commands which are
scored on a three point scale and are
presented in conjunction with gestural
cues. The SIB also allows for non-
verbal and partially correct responses.
The six major subscales are attention,
orientation, language, memory, visuo-
spatial ability, and construction.
Overall scores range from 0-1000 with
positive scores indicating clinical
improvement
Syndrom Kurz Test A psychometric test battery for the This test has shown good test-retest
(SKT) assessment of memory and attention. reliability. Correlations with other
The SKT consists of nine 1 minute cognitive measures support its validity
subtests that are partly speed oriented as a cognitive outcome measure for
and partly span orientated: scaled AD.
subtest scores are aggregated to an
SKT total status score ranting from 1
(very good) to 27 (very poor).
Ten Point Clock This is a screening test for dementia in This test has been shown to be both
Drawing Test particular for assessing visuospatial reliable and valid and is simple and
and executive functions. Patients have easy to administer with good
to drawn in the numbers of digits sensitivity and specificity.
placed in a pre drawn circle.
Confidential material removed PenTAG 2010
- 5 -
AChEIs & memantine for Alzheimer's Appendices

Type Construct measure and scoring Critical appraisal
Trail Making Test Assesses speed of visual search, Reliability is reported to be higher for
(TMT) attention, mental flexibility and motor part A than for part B, which requires
function. The test has two parts: A) more information-processing ability
drawing a line linking numbers in and is more sensitive to brain damage.
sequence and B) drawing a line linking Reliability is restricted due to the use
letters in sequence. The reviewer calls of time scores rather than both error
any mistakes to the attention of the counts and time scores, since error
participant, and these must be correction may take longer in some
corrected before progressing. The participants than others. Scores are
score is the time taken to successfully strongly affected by the participant‟s
complete a test. education level.
Wechsler logical This test is one of 13 subtests of the Test-retest reliability and concurrent
memory test Wechsler Memory Scale-Revised. The validity with a verbal learning test are
first subtest is for screening purposes, adequate for the whole WMS-R test.
and the other 12 are grouped into five Level of education affects a
separate memory areas. The test participant's score. Normative data for
manual provides guidelines for scoring those aged 75 and over is lacking. The
and weighting, and provides norms for score is more heavily influenced by
individuals aged 16-74 with verbal memory performance than by
information about significant other memory components.
differences between any two scores.
Functional and quality of life outcome measurement scales
Type Construct measure and scoring Critical appraisal
Alzheimer‟s This rating scale is a 23-item assessment of The ADCS-ADL is a structured
Disease ADLs that is scored from 0 (greatest questionnaire originally created to
Cooperative impairment) to 78. It evaluates Activities of assess functional capacity over a
Study-Activities of daily living. broad range of severity of dementia.
Daily Living The ADAS-ADL19 is a subset of the
original inventory and focuses on
ADCS-ADL
items appropriate for the assessment
of later stages of dementia. The
sensitivity and reliability of this
modification has been established.
Alzheimer's Scale consists of 10 items for instrumental Full assessment of psychometric
Disease ADL: ability to use the telephone, properties not yet published. Has
Functional performing household tasks, using face validity for those with mild-
assessment and household appliances, handling money, moderate AD.
Change Scale shopping, preparing food, ability to get The ADL items chosen for this scale
(ADFACS) around both inside and outside the home, have been demonstrated to be
pursuing hobbies and leisure activities, sensitive to change over 12 months,
handling personal mail, grasping situations correlate well with MMSE scores, and
or explanations. Scale has a range of 0 to have good test-retest reliability
54 where lower scores correspond to better (although several questions have
function. Test takes approximately 20 been modified in the scale).
minutes to complete.
Confidential material removed PenTAG 2010
- 6 -
AChEIs & memantine for Alzheimer's Appendices

Type Construct measure and scoring Critical appraisal
Behavioural Consists of 35 items (scored 0, 1, or 2) Unable to source data on reliability
Rating Scale for assessing observable aspects of cognition, and validity.
Geriatric Patients function and behaviour. A high score
(BGP) indicates worse function.
Bristol Activities Caregiver assessment of 20 ADLs. Designed specifically for use with
of Daily Living Categories included are food, eating, patients with dementia. Face validity
scale (BADL) drinks, drinking, dressing, hygiene, teeth, was measured by asking carers
bath, toilet, transferring, mobility, orientation whether items were important, and
to time and space, communication, construct validity was confirmed by
telephone, housework/gardening, shopping, principal components analysis.
finances, hobbies, and transport. Scores Concurrent validity was assessed by
range from 0 - 60 with higher scores observed performance, the test has
indicating better function. good content validity, and there is
good test-retest reliability. The test is
shown to correlate well with
performance ADLs and tests of
cognitive function.
Caregiver-rated A modified Crichton Geriatric Rating Scale Reliability demonstrated. Unable to
Modified Crichton (CGRS). This a seven-item scale using a source data on validity.
Scale (CMCS) Likert-type scoring method. Questions
include comprehension to time and place,
carrying out conversation, cooperation,
restlessness, dressing, social activities and
leisure. Negative change relates to clinical
improvement.
Disability This rating scale is a 46-item structured The DAD scale demonstrates a high
Assessment interview or questionnaire for the caregiver degree of internal consistency and
for Dementia that is scored from 0 to 100 (least excellent interrater and test-retest
(DAD) impairment). It evaluates ADLs and takes reliability. Full details of concurrent
approximately 20 minutes to complete. It is and construct validity not yet
based on a recognised conceptual definition published.
of disability from the WHO
Functional Assesses the magnitude of progressive FAST has been shown to be a
Assessment functional deterioration in patients with reliable and valid assessment
Staging scale dementia by identifying characteristic technique for evaluating functional
(FAST) progressive disabilities. Seven major stages deterioration in AD patients
range from normal (stage 1) to severe throughout the entire course of the
dementia (Stage 7). illness. Because the elements of
functional capacity incorporated in
FAST are relatively universal and
readily ascertainable, as well as
characteristic of the course of AD,
FAST can serve as a strong
diagnostic and differential diagnostic
aid for clinicians.
General Health GHQ-30 The GHQ is a self-report GHQ-30 is based on Medical
Questionnaire psychiatric screening test, and items Outcomes Study Short Form-36,
(GHQ-30) include questions on: depression and which is extensively validated
Confidential material removed PenTAG 2010
- 7 -
AChEIs & memantine for Alzheimer's Appendices

Type Construct measure and scoring Critical appraisal
unhappiness, anxiety and felt psychological
disturbance, social impairment, and
hypochondriasis. Participants rate
themselves on a four-point severity scale,
according to how they have recently
experienced each GHQ item: better than
usual, same as usual, worse than usual, or
much worse than usual. Normally each
item is scored either 0 or 1, depending on
which severity choice is selected. Individual
items are summed to give the total score.
Instrumental For women, the set of behaviour assessed The IADL is a very frequently used
Activities of Daily include telephoning, shopping, food and often cited instrument for
Living (IADL) preparation, housekeeping, laundering, use assessing the instrumental
of transport, use of medicine and ability to competence of elderly patients. The
handle money. For men, the areas of food scale is well anchored from a
preparation, housekeeping and laundering theoretical point of view and the
are excluded. behaviours that are included are
Each of the behavioural areas is given a likely to be affected in the first stages
score of 0 or 1, leading to an overall score of dementia.
that ranges from 0 to 8 for women and from
0 to 5 for men.
The Interview for The IDDD measures functional disability in This scale appears to be appropriate
Deterioration in self-care (16 items such as washing, to assess community-living patients
Daily Living in dressing and eating) and complex activities with mild and moderate levels of
Dementia (IDDD) (17 items such as shopping, writing, and dementia. It assesses a substantial
answering the telephone) proportion of complex activities likely
Severity of impairment is rated on a 7-point to be affected during the first stages
scale, where 1-2=no or slight impairment, of the AD. The number of non-
redundant items in the scale is
3-4=mild impairment, 5-6=moderate
viewed positively since it may
impairment, 7=severe impairment, giving a
increase the sensitivity of the tool.
total range score of 22-231.
Empirical info on the testing of the
IDDD and its measurement
properties is seriously lacking.
Physical Self- Measured through competence of 6 Brief assessment of activities of daily
Maintenance behaviours: toileting, feeding, dressing, living. Theoretically well grounded, it
Scale (PSMS) grooming, locomotion and bathing. It can be has been proven useful for evaluation
completed by untrained staff based on of institutionalised elderly but has a
information from subjects, caregivers, ceiling effect for those living in the
friends etc. Each behavioural area is given community. Testing of psychometric
a score of 1 or 0, with over score ranging properties is incomplete.
from 0 to 6. Using Guttman scaling, each
scale point has 5 descriptive scale points.
The Progressive PDS examines activities of daily living and This scale has been shown to be
Deterioration instrumental activities of daily living. sensitive to three severity stages of
Scale (PDS) Examples are: extent to which a patient can dementia although some debate
leave the immediate neighbourhood, use of whether the content is adequate to
familiar household implements, involvement assess those with moderately-severe
Confidential material removed PenTAG 2010
- 8 -
AChEIs & memantine for Alzheimer's Appendices

Type Construct measure and scoring Critical appraisal
in family finances, budgeting. AD. The scale was systematically
Each question is scored by measuring the developed and tested on a fairly large
distance along the line on a scale from 0 to sample of AD patients (although the
100, with higher scores reflecting better mean age of the final test group was
functionality. A composite score is derived only 69.5 years).
from averaging across the items for a Test-retest reliability was determined
maximal score of 100. in 123 patients, giving stage
The scale is sometimes classified as a correlations (rs) of 0.889 for early AD
measure of quality of life. (14 participants), 0.775 for 44 middle
stage participants and 0.775 for 65
late stage participants. A moderate
degree of correlation has been
demonstrated between PDS and
ADAS-cog scores (rp= -0.57 to -
0.64).
There is considerable reduplication
within the scale – 4 questions relate
to handling finances but there are no
items pertaining to basic activities
such as washing, dressing and
toileting. The scale is therefore not
thought to have adequate content to
assess people with moderately
severe AD as it does not assess the
wide range of daily living skills
affected at different stages of the
disease. There are high levels of
between and within patient variability
(in the order of 12 points) which may
make it less suited to detect
differences over short time periods.
QOL (patient and This assessment was a 7-item patient-rated This instrument has not been
caregiver scales) scale evaluating the patients perceptions of validated in patients with Alzheimer's
their well-being in terms of relationships, disease but was selected because no
eating and sleeping, and social and leisure QOL instrument has been validated
activities. The tests is conducted by in this population.
interview. Scored on an analogue scale
between 0 (worst quality) to 50 (best
quality).
Unified Activities All self-care and mobility variables The psychometric properties of this
of Daily Living commonly used to assess patient‟s scale, resulting from the combination
Form (Unified functional status. of existing evaluations, have not
ADL) A 20-item scale was produced. The need been published.
for assistance is scored for every item, on a
10-point scale.
Confidential material removed PenTAG 2010
- 9 -
AChEIs & memantine for Alzheimer's Appendices

Behaviour and mood outcome measurement scales
Type Construct measure and scoring Critical appraisal
Behavioural A measure of the severity of behavioural The BEHAVE-AD has been
Pathology in symptoms in AD. It consists of 25 symptoms shown to be reliable and valid.
Alzheimer‟s group onto seven categories. Each symptom is
Disease rating scored on the basis of severity on a four point
scale (BEHAVE- scale.
AD)
Behavioural A 35 item rating scale more commonly used in No information about the
Rating Scale for European trials. reliability or validity of this scale
Geriatric patients was found.
(BGP)
NOSGER - Contains 30 items of behaviour, each rated on This scale has been validated,
Nurses a 5-point scale according to frequency of and has high inter-rater and test-
Observation occurrence. Item scores are summarized into 6 retest reliability. The test
Scale for Geriatric dimension scores (memory, instrumental correlates well with clinician's
Patients activities of daily life, self-care, mood, social global rating of change.
behaviour, and disturbing behaviour).
Neuro-psychiatric Currently evaluates 12 items: delusions, Content validity has been
Inventory (NPI) hallucinations, dysphoria, anxiety, agitation, established, reliability and validity
euphoria, apathy, irritability, disinhibition, are satisfactory. Limitations
aberrant motor behaviour, night-time behaviour included: poor description of
and changes in appetite/eating behaviour. appraisal period for behavioural
Psychometric properties were established on symptoms; no justification for
first 10 items. Total score for each domain is scoring system; and, inter-rater
calculated by multiplying frequency rating by reliability was poorly deserved.
severity rating, adding domain scores to get a
total score. Higher scores represent more
problems. Maximum scores is 12 per domain,
with either 10 or 12 domains assessed.
Confidential material removed PenTAG 2010
- 10 -