Post on 31-Mar-2019
Technical Appendix
Subjects
Further details of inclusion and exclusion criteria are detailed below:
Inclusion Criteria
1. A clinical diagnosis of possible Alzheimer's disease in accordance with the National Institute of
Neurological and Communicative Diseases and Stroke/Alzheimer's Disease and Related
Disorders Association (NINCDS-ADRDA) criteria, with radiological (Magnetic Resonance Imaging
[MRI] or Computed Tomography [CT]) evidence of significant cerebrovascular disease (CVD),
assessed within the last 12 months
2. Male or female between 50 and 80 years of age inclusive, at the time of signing the informed
consent.
3. Subject has a documented history of at least 6 months of ongoing Alzheimer's disease therapy
(AChEIs and/or memantine) with stable dosing for at least the last 2 months (and with no intent
to change for the duration of the study).
4. Mini-Mental Status Examination (MMSE) score between 20 and 26 at Screening.
5. Clinical Dementia Rating Scale (CDR) score of 0.5 or 1.0 at Screening.
6. A female subject is eligible to participate if she is of non-childbearing potential defined as pre-
menopausal females with a documented tubal ligation or hysterectomy; or postmenopausal
defined as 12 months of spontaneous amenorrhea; or of child-bearing potential and agrees to
use acceptable contraception methods
7. Subject has provided full written informed consent prior to the performance of any protocol-
specified procedure; or if the subject is unable to provide informed consent due to cognitive
status, the subject will provide assent and full written informed consent will be provided by a
legally acceptable representative
8. The subject has a dedicated caregiver who is willing to supervise participation in the study
9. In the opinion of the investigator, the subject has the ability to comply with study procedures
(cognitive and other testing) and is fluent in the language used for the administration of the
cognitive tests.
Exclusion Criteria
1. History and/or evidence of any other CNS disorder that could be interpreted as a cause of
dementia: e.g. structural or developmental abnormality, epilepsy, infectious, degenerative or
inflammatory/demyelinating CNS conditions such as, Parkinson's disease and frontotemporal
dementia
2. History of significant psychiatric illness such as schizophrenia or bipolar affective disorder that in
the opinion of the Investigator would interfere with participation in the study; major depressive
disorder (according to DSM-IV) in the past year; current active depression requiring initiation of
treatment (or is believed to account for substantial degree of cognitive impairment)
3. Evidence of the following disorders: current vitamin B12 deficiency, positive syphilis serology
(unless neurosyphilis was ruled out) or active thyroid dysfunction (particularly suggestive of
hypothyroidism), including abnormally high or low serum levels of thyroid stimulating hormone
(TSH), where this is thought to be the cause of, or to contribute to the severity of the subject's
dementia.
4. History of alcohol or other substance abuse, according to the DSM-IV criteria, or recent or
remote history of the same if that could be a contributing factor to dementia.
5. History of intra cerebral haemorrhage due to any of the following causes: cerebral amyloid
angiopathy, uncontrolled hypertension, cerebral arteriovenous malformation, coagulopathy,
CNS vasculitis or any other condition that the investigator and/or medical monitor considers as
a relevant risk factor for intracerebral haemorrhage
6. Recent (i.e.,<6 months from Screening Visit) cardiovascular event defined as:
a. ST-elevation MI or non-ST-elevation MI, confirmed by cardiac enzyme elevation and ECG
changes
b. coronary revascularization (percutaneous coronary intervention or coronary artery bypass
graft )
c. stroke of any etiology
d. resuscitated sudden death
e. prior carotid surgery or stenting procedure
7. Poorly controlled hypertension despite lifestyle modifications and pharmacotherapy (either
systolic blood pressure >160mmHg or diastolic blood pressure >110mmHg)
8. QTcB interval >450 msec; or QTcB > 480 msec in subjects with bundle branch block based on
ECG assessment at the Screening visit.
9. HbA1c >12.0 at Screening, or uncontrolled diabetes in the opinion of the investigator.
10. History of glaucoma or any other findings in the baseline eye exam that, in the opinion of the
investigator, would exclude the subject from participation in the study.
11. History of adult asthma (or reactive airway disease) manifested by bronchospasm in the past 6
months, or currently taking regular anti-asthmatic medication(s).
12. Previous history of anaphylaxis, severe allergic reaction or history of hypersensitivity to any of
the components of the formulation.
13. Significant abnormalities on clinical chemistry, haematology or urinalysis at Screening, including
clinically significant anaemia.
14. History of chronic viral hepatitis (including presence of B surface antigen or hepatitis C
antibody), or other chronic hepatic disorders.
15. Abnormal Screening blood tests exceeding any of the limits defined below:
a. Alanine transaminase (ALT) or aspartate transaminase (AST) >1.5 x the upper limit of normal
(ULN)
b. Alkaline phosphatase (AP) and bilirubin >1.5X ULN (isolated bilirubin >1.5x ULN is acceptable
if bilirubin is fractionated and direct bilirubin <35%).
c. Calculated creatinine clearance < 30 ml/min (per Cockcroft & Gault) at Screening
16. Other clinically significant abnormality on physical (including neurological), laboratory or ECG
examination that could be detrimental to the subject in the opinion of the Investigator or could
compromise the integrity of the study.
17. Planned major surgery within the study period.
18. Use of systemic steroids or other immunosuppressants within the last 30 days prior to
screening.
19. Current treatment with barbiturates, MAO inhibitors, butyrophenones, phenothiazines and
other "conventional" antipsychotic within 30 days or 5 half-lives prior to Screening, whichever is
longer.
20. Treatment with antidepressants, (other than MAO inhibitors), thyroid hormones, atypical
antipsychotics (e.g. risperidone), benzodiazepines and other sedatives / hypnotics unless
prescribed at a stable dose for at least 2 months prior to Screening. Note: Benzodiazepines or
other sedatives/hypnotics (including antihistamines) with half-life less than 6 hours can be
taken on a prn (as needed) basis but must not be taken within 5 half lives prior to cognitive
testing.
21. Current treatment with known potent inhibitors of CYP3A4 (e.g. ketoconazole, rifampin,
modafinil).
22. Current treatment with known potent Pgp inhibitors (itraconazole, ketoconazole, cyclosporin,
loperamide, diltiazem, verapamil, spironolactone, quinidine, bepridil, quinine, carvedilol)
23. Cognitive tasks prescribed for cognitive rehabilitation and performed under medical supervision
in the 6 months prior to screening and/or during study
24. Investigational medications or devices including symptomatic AD treatment during the 60 days
prior to the Screening visit, or within 5 half-lives of use of the investigational drug prior to the
Screening Visit, whichever is longer.
25. Participation in another investigational drug (with the exception of anti-amyloid monoclonal
antibodies [mAbs]) or device study where subject was treated chronically (i.e. > 1 single dose)
with a study agent intended to impact AD progression during the 12 months prior to the
Screening visit.
a. Subjects who participated in an investigational drug study that involved chronic dosing with
a monoclonal antibody at any time in the past are excluded from this study, unless it is
known that they received placebo during the previous study.
b. Subjects who participated in previous single-dose studies of anti-amyloid mAbs will be
permitted provided the subject's dose of the mAb is at least 5 half-lives removed; the
subjects did not experience any moderate adverse events classified as possibly drug-related
or any serious adverse event during that study; the subject did not drop out of the previous
study (i.e. completed all safety assessments)
26. Subjects, who in the investigator's judgement, pose a significant suicide risk (e.g. history of
suicidal behaviour in the last 6 months and/or any suicidal ideation of type 4 or 5 on the C-SSRS
in the last 2 months).
27. Subject or caregiver is an immediate family member or employee of the participating
Investigator, any of the participating site staff or GSK staff.
28. Any contraindication to lumbar puncture or insertion of CSF catheter, including but not limited
to
a. Thrombocytopenia or other coagulation disorders (including subjects receiving coumarin-
derived anti-coagulants or low-molecular-weight heparin).
b. The presence of cutaneous or soft tissue infection overlying or adjacent to the site of lumbar
puncture.
c. Previous spinal surgery that could complicate access to the subarachnoid space.
d. Suspicion of increased intracranial pressure due to a cerebral mass.
Study procedures, outcomes and assessments
Assay Details
Plasma LpPLA2 activity was assessed via a colorimetric activity method by Quest Diagnostics, UK.
CSF LpPLA2 activity was assessed via a radiometric 3H-PAF assay method by Wuxi AppTec Co. Ltd,
China.
CSF A1-42, CSF A1-40, CSF t-tau and CSF 181P-tau were assessed by ELISA using the INNOTEST
(Innogenetics) platform at Sahlgrenska University Hospital, Sweden.
CSF NF-L was assessed by ELISA at Sahlgrenska University Hospital, Sweden.
Albumin Quotient was assessed by Nephelometry at Sahlgrenska University Hospital, Sweden.
Further Details of Cognitive Assessments
A number of computerised, and pen-and-paper neuropsychological tests described below were
performed at screening, baseline, week 12 and week 24 in order to assess the impact of treatment
on cognition, particularly verbal learning, attention, episodic memory, working memory and
executive function. Wherever possible, the tests were performed at approximately the same time of
day on each assessment day and took approximately 25-30 minutes to complete. All cognitive
assessments were administered by a healthcare professional experienced in administering
neuropsychological tests, who had successfully completed the study specific rater training co-
ordinated by CogState. Wherever possible, assessment was performed at approximately the same
time of day and by the same rater at each visit.
Episodic Memory
International Shopping List Test (ISLT) Immediate Recall: This is a computer controlled verbal
learning test, in which subjects were read a list of 12 words. After the 12 words had been read, the
subject tried to recall as many of the words as they could. When they could recall no more words the
same list was read a second time with the words in the same order. The process of reading the list
and waiting for responses occurred three times. The primary outcome measure was the number of
correct words recalled.
International Shopping List Test Delayed Recall: At the completion of the computerised battery the
subject was asked to recall as many of the items as they could from the shopping list. This provided a
measure of delayed recall. The primary outcome measure was the number of correct words recalled.
Attention
Identification task: This is a measure of choice reaction time and has been shown to provide a valid
assessment of visual attention. In this task an event (a card turning face up) occurred in the centre of
the computer screen and the subject had to decide “YES” or “NO” as to whether this event met a
predefined and unchanging criterion (is the colour of the card red?). The software measured the
speed and accuracy of each response. Performance speed was the primary outcome measure.
Trails A: Subjects had to connect 25 targets containing a number in the correct sequence as quickly
as possible (i.e. 1-2-3 etc). The targets were distributed in a random fashion across a page. The test
was a measure of psychomotor speed and attention. Completion time was the primary outcome
measure.
Executive Function / Working Memory
Controlled Oral Word Association: This task was a brief (<4 minutes) measure of language fluency,
planning and working memory. Subjects were instructed to generate as many words as they could
think of beginning with a specific letter in one minute. They were then requested to do exactly the
same for two furthers letters. Subjects were requested not to say proper nouns, words beginning
with same stem but with a different ending and not to repeat themselves. Stimulus letters for this
test were selected on the basis that they had high frequency scores when tested in normal healthy
volunteers and as such there were some variations in the letters used from country to country. Each
correct response scores one point. Performance for all three letters was summed to yield a total
number of acceptable responses and this was the primary outcome measure.
Category Naming: This test measures semantic fluency, planning and working memory. Subjects
were required to generate as many exemplars of the category ‘Animals’ as they could in one minute.
The animals mentioned had to be real animals but could be living or extinct for the purposes of this
test. The rules of the Category Naming test required that credit was not given for imaginary animals
or for repetitions. Each correct response scored one point and the key metric for this test was the
total number of acceptable responses.
One-back: This is a measure of working memory where the subject was shown a single stimulus in
the centre of the computer screen (a card turned face up). They had to decide “YES” or “NO” as to
whether the current card matched the card that had been seen on the immediately previous trial.
The software measured the speed and accuracy of each response. Accuracy of performance was the
primary outcome measure.
Trails B: Like Trails A, subjects had to connect 25 targets in the correct sequence as quickly as
possible. However, these targets contained either a letter of the alphabet or a number. The subjects
had to connect the targets by alternating between letters and numbers (i.e. A-1-B-2-C-3 etc). The
targets were distributed in a random fashion across a page. In addition to motor speed, visual
scanning, and visual-motor integration, this test required attention and cognitive flexibility. The
number of seconds required to complete the task was measured and was the primary outcome
measure.
Go No-Go task: In this computerised task stimuli (playing cards) were presented in a continuous
stream and participants performed a binary decision (was it black?) on each stimulus. One of the
outcomes required participants to make a motor response (go), whereas the other required
participants to withhold a response (no-go). ‘Go’ events typically occurred with higher frequency
than ‘no-go’ events. Accuracy and reaction time were measured for each response. The primary
outcome measure was the speed of performance.
Calculation of CogState composite scores:
Three Composite Scores were derived from the nine Task Scores. The Composite Score was
calculated by taking the mean of specific Standardised Task Scores as described below.
Executive function/working memory (EF/WM) composite score One Back Memory Controlled Oral Word Association Category Fluency Go/No Go Task Trails B
In order to calculate the EF/WM composite score item scores must have been available for at least 3
of the 5 above tasks. A further requirement was that there must have been at least one non-missing
response from pen and paper tasks (COWAT/Trails B/Category Fluency) and one non-missing
response from computerised tasks (One back memory or Go/No go). Assuming these criteria were
met the composite score was calculated as the mean of the available item scores. If these criteria
were not met then a subject had a missing EF/WM composite score).
Attention composite score Identification Trail Making Test A
Both items needed to be present for the attention composite score to be calculated.
A Total Composite Score was also derived from the nine item Scores. The Total Score was calculated
by taking the mean of all nine Task Scores.
When calculating the totals composite, in situations where there were missing item scores, it was
required that the following minimum data were available:
At least 5 non missing item scores, ISLT non missing At least one non missing item for either Identification or Trails A At least one of these non-missing items coming from the any of the COWAT/Trails
B/Category Fluency item scores At least one non-missing score for One back memory or Go/No go.
Assuming these criteria were met, the composite score was then calculated as the mean of the
available item scores. In situations where the criterion above was not met the composite total score
was set to missing.
Additional details for the Bayesian Analysis:
Bayesian methods of analysis were used to estimate the posterior distribution for the treatment
difference for CSF and CogState parameters.
The Bayesian analysis of the CSF parameters, which were collected at a single timepoint post
baseline, was based on a non-informative prior, using the prior statement within the mixed
procedure in SAS (with the default Jeffreys prior option). Posterior probabilities and 95% credible
intervals for the treatment difference at End of Study being greater than pre-defined cut-offs were
produced for the change from baseline in CSF parameters for the ITT population.
For the Bayesian analysis of the Cogstate parameters, which were collected at two post baseline
timepoints, Markov chain Monte Carlo (MCMC) methods were used as a means of sampling from the
posterior distribution of interest from an observed cases analysis, which incorporated the repeated
measures nature of the collected data. The MCMC procedure in SAS used an uninformative inverse-
wishart prior for the covariance and flat priors for the linear parameters. MCMC algorithms, using
Gibbs sampling, enabled calculation of the posterior distribution of the treatment difference.
Posterior probabilities and 95% credible intervals were also calculated for the effect size being
greater than pre-defined cut-offs of 0, 0.15 and 0.3 for the change from baseline in CogState working
memory / executive function in the ITT population.
Supplementary Materials3.1 Patient Disposition and DemographicsConcomitant Medications
Supplementary Table 1 Summary of Alzheimer’s disease concomitant medications taken during the treatment phase (ITT Population)
Placebo(N=61)
Rilapladib 250mg(N=60)
Any medication 61 (100) 60 (100)Donepezil / Donepezil hydrochloride 28 (46) 41 (68)Rivastigmine / Rivastigmine Hydrogen Tartrate 20 (33) 14 (23)Memantine Hydrochloride 8 (13) 7 (12)Galantamine / Galantamine Hydrobromide 10 (16) 2 (3)
ITT, Intent-to-treatNon-AD concomitant medications were taken by 92% subjects in the placebo group and 98% subjects in the rilapladib group. The most commonly used treatments were simvastatin and acetylsalicylic acid.
3.2 Pharmacokinetics and PharmacodynamicsPlasma LpPLA2 Activity
Supplementary Table 2. Summary of Percentage Inhibition in Plasma Lp-PLA2 Activity by Visit (ITT Population)
Visit Placebo(N=61)
Rilapladib 250mg(N=60)
Week 1 n 56 53Mean -2.39 80.17
SD 23.345 18.389Median 2.89 85.13
Min. -109.3 -16.0Max. 23.8 92.2
Week 12 n 52 52Mean -1.30 78.09
SD 24.294 25.612Median -0.98 87.38
Min. -129.9 -35.9Max. 60.0 93.7
Week 24 n 53 47Mean -3.86 83.59
SD 30.228 10.483Median -2.36 86.56
Min. -141.1 43.5Max. 70.4 96.3
Follow-up n 1 2Mean -10.47 44.61
SD 30.010Median -10.47 44.61
Min. -10.5 23.4Max. -10.5 65.8
% inhibition=(change from baseline in Plasma LpPLA2/baseline LpPLA2)*-100LpPLA2, lipoprotein phospholipase A2; ITT, intent-to-treat; SD, standard deviation; min, minimum; max, maximum
CSF LpPLA2 ActivitySupplementary Table 3. Summary of Change from Baseline in CSF Lp-PLA2 Activity
(mcmol/min/L) at End of Study
Visit Placebo(N=61)
Rilapladib 250mg(N=60)
Baseline n 59 60Mean 1.361 1.852
SD 4.6064 6.2783Median 0.770 0.695
Min. 0.21 0.27Max. 36.07 35.52
Change to End of Study* n 50 48Mean 0.026 -0.464
SD 7.2657 5.4263Median 0.010 -0.060
Min. -35.74 -34.64Max. 36.16 13.90
*End of Study includes the final evaluable post baseline assessments including those conducted during the follow up period.CSF, cerebrospinal fluid, LpPLA2, lipoprotein phospholipase A2; SD, standard deviation; min, minimum; max, maximum
3.3 Efficacy
Aβ Biomarkers
Supplementary Table 4. Per-protocol and Sensitivity analyses of change from baseline at week 24 (ANCOVA) in CSF Aβ1-42 and Aβ1-40 (ng/L)
Analysis/population Treatment n Adjusted Mean (SE)
Difference vs, placebo†, (95% CI)
Per protocol PlaceboRilapladib 250mg
5043
-6.1 (18.95)43.4 (20.45) 49.5 (-6.1, 105.0); p=0.080
Aβ1-42 ITT Excluding Outliers 1, 2 PlaceboRilapladib 250mg
5144
-3.3 (12.41)5.3 (13.37) 8.6 (-27.8, 44.9)
Aβ1-40 ITT Excluding Outliers 1, 2 PlaceboRilapladib 250mg
5144
-31.4 (158.15)-122.2 (170.29) -90.7 (-552.9, 371.4)
Aβ1-42 ITT Excluding Early Withdrawal and Follow-Up Data
PlaceboRilapladib 250mg
4743
-15.6 (19.94)40.8 (20.85) 56.4 (-1.2, 114.0); p=0.055
Aβ1-40 ITT Excluding Early Withdrawal and Follow-Up Data
PlaceboRilapladib 250mg
4743
-164.0 (191.04)-427.0 (199.73) -263.0 (-812.4, 286.5)
ITT= intent-to-treat; SE= standard error†Difference in adjusted least square means is shown (Rilapladib minus Placebo)1Analyses were conducted post-hoc26 subjects were removed from this analysis on the basis of having studentised residuals >=+/-2 for the primary ITT analysis
Cognition
Supplementary Table 5. Sensitivity analysis of change from baseline at week 24 in CogState Executive function/working memory score (ITT Population unless stated)
Analysis Method/population
Treatment n Adjusted Mean (SE)
Difference vs, placebo†,
(95% CI)
Effect Size (ES) (95% CI)
Per protocol (MMRM)1
PlaceboRilapladib 250mg
53 42
-0.137 (0.0517) 0.016 (0.0577) 0.153 (-0.001, 0.307) 0.408 (-0.003, 0.818)
Completers(ANCOVA) 2
PlaceboRilapladib 250mg
5648
-0.151 (0.0498)0.034 (0.0539) 0.185 ( 0.039, 0.331) 0.495 ( 0.104, 0.886)
Multiple Imputation (J2R) 3
PlaceboRilapladib 250mg
57 54
-0.153 (0.0482)-0.003 (0.0571) 0.150 ( 0.002, 0.299) 0.383 ( 0.004, 0.762)
All inclusive data (MMRM) 4
PlaceboRilapladib 250mg
5652
-0.153 (0.0538)-0.032 (0.0555) 0.121 (-0.033, 0.274) 0.301 (-0.081, 0.683)
ITT= intent-to-treat; SE= standard error †Difference in adjusted least square means is shown (Rilapladib minus Placebo).1 Per protocol population defined as subjects in the ITT population who were not major protocol deviators. MMRM model fitted as per primary analysis.2 Analysis of covariance (ANCOVA) of all subjects who provided week 24 CogState data. Analysis model adjusts for baseline and treatment.3 Multiple imputation method using a Jump to reference assumption; Subjects who withdraw from Rilapladib are assumed to ‘switch’ treatment to placebo, therefore missing values for these subjects are imputed using the mean placebo profile..4 Analysis removes pre-specified requirements in the reporting and analysis plan around minimum amount of data required in order to calculate the EF/WM composite scores.
3.4 Safety and Tolerability
Supplementary Table 6. Summary of Serious Adverse Events by Treatment Phase (Safety Population)
Serious Adverse Event Placebo(N=62)
Rilapladib 250mg(N=61)
Any on-treatment Serious Adverse Event, n (%) 5 (8) 8 (13)Fall 0 1 (2)Femoral neck fractureFractured coccyx
0 1 (2)1 (2) 0
Scapula fracture 0 1 (2)Sternal fracture 1 (2) 0Colitis microscopic 0 1 (2)Inguinal hernia 1 (2) 0Large intestine perforation 1 (2) 0Cerebral haemorrhage 0 1 (2)Complex partial seizures 0 1 (2)Dizziness 0 1 (2)Upper respiratory tract infection 1 (2) 0Vestibular neuronitis 1 (2) 0Bradycardia 0 1 (2)Any Serious Adverse Event in follow–up event, n (%) 1 (2) 1 (2)Pneumonia 1 (2) 0Leukaemia 0 1 (2)Pulmonary embolism 1 (2) 0