1

Comparative Safety and Effectiveness of Cognitive Enhancers for the Treatment of Alzheimer’s Disease: A Rapidly Updated Systematic Review and Network Meta-analysis FINAL REPORT Andrea C. Tricco, Huda M. Ashoor, Patricia Rios, Jemila Hamid, John D. Ivory, Paul A. Khan, Fatemeh Yazdi, Charlene Soobiah, Marco Ghassemi, Erik Blondal, Joanne Ho, Sharon E. Straus

August 5th, 2015 Disclaimer: The information in this report is a summary of available material and is designed to give readers (health systems stakeholders, policy and decision makers) a starting point in considering currently available research evidence. Other relevant scientific findings may have been reported since completion of the review. This report is current to the date of publication and may be superseded by an updated publication on the same topic. This is a preliminary report and will be modified prior to final publication. You should consult other sources in order to confirm the currency, accuracy and completeness of the information contained in this publication and, in the event that medical treatment is required you should take professional expert advice from a legally qualified and appropriately experienced medical practitioner.

30 Bond Street, Toronto ON, M5B 1W8

www.odprn.ca

[email protected]

2

Executive Summary

Summary This rapid review and network meta-analysis was conducted to update an existing review of the safety and efficacy of cognitive enhancers for patients with Alzheimer’s Disease. A network meta-analysis found that donepezil improves cognition and global status, rivastigmine improves cognition and decreases risk of mortality, and galantamine improves global status. All cognitive enhancers improved cognition for mild-to-moderate AD patients and donepezil improved cognition for severe AD patients. However, all three caused increased risk of gastrointestinal harm and rivastigmine causes increased risk of headaches. The rivastigmine patch was similar in efficacy to oral rivastigmine but caused less nausea. There were no differences between agents in the risk of bradycardia, falls or overall serious adverse events. Implications As this is a rapidly updated review, results should be interpreted with caution. However these findings demonstrate the need for patient decision aids to help individuals with AD and their families to choose the right medication for their treatment goals. Future research in this area will focus on analysing individual patient data in a network meta-analysis. What is the current practice in treating Alzheimer’s Disease with cognitive enhancers? • Patients with AD are often treated with cognitive enhancers, including donepezil, rivastigmine, galantamine and memantine; however, it is not clear which of these therapies or combinations of therapies are safest and most effective Objective • The objective was to rapidly update an existing systematic review and network meta-analysis (NMA) examining the comparative safety and efficacy of cognitive enhancers for patients with Alzheimer’s Disease How was the study conducted? • A librarian in the group updated the review by searching 6 electronic databases from September 2011 until January 2015 • We included randomized and non-randomized studies that examined cognitive enhancers (donepezil, rivastigmine, galantamine, and memantine) alone or in combination compared to each other or placebo • The outcomes of interest were: cognition, function, behaviour, global status, mortality, and harms (nausea, vomiting, diarrhea, bradycardia, headache, falls, and all serious adverse events (SAEs)) • Screening of literature search results, data abstraction, and risk of bias appraisal were completed independently by two reviewers • Random-effects Bayesian NMA was conducted for outcomes with a connected network of treatments What did the study find? • 186 studies including 106 RCTs, 20 non-randomized clinical-trials, 8 cohort studies, and 53 companion reports were included

Ontario Drug Policy Research Network

3 •

• • • •

Donepezil and rivastigmine improved cognition compared to placebo, but only donepezil showed a clinically important difference on the ADAS-cog scale. All cognitive enhancers improved cognition compared to placebo in patients with mild-to-moderate AD and donepezil improved cognition compared to placebo in patients with severe AD. All cognitive enhancers improved global status compared to placebo and galantamine demonstrated a clinically meaningful effect compared to placebo, donepezil, rivastigmine, and donepezil+memantine; no differences were observed between the agents for function or behaviour Donepezil, rivastigmine, and galantamine all caused more gastrointestinal side effects compared to placebo; no differences between agents were observed for SAEs, bradycardia or falls Rivastigmine decreased the risk of mortality compared to placebo and memantine but caused more headaches compared to placebo Rivastigmine patch demonstrated similar efficacy in improving cognition but caused significantly less nausea than oral rivastigmine

Ontario Drug Policy Research Network

4

Table of Contents

Executive Summary....................................................................................................................................... 2 Summary ............................................................................................................................................... 2 Implications ........................................................................................................................................... 2 RATIONALE .................................................................................................................................................... 7 METHODS ...................................................................................................................................................... 7 Eligibility criteria ....................................................................................................................................... 7 Information sources and literature search ............................................................................................... 7 Study selection process............................................................................................................................. 8 Data items and data abstraction process ................................................................................................. 8 Synthesis of included studies .................................................................................................................... 8 RESULTS ........................................................................................................................................................ 9 Literature search ....................................................................................................................................... 9 Study and patient characteristics ........................................................................................................... 10 Risk of bias and methodological quality appraisal .................................................................................. 10 Efficacy outcomes ................................................................................................................................... 10 Cognition – Overall MMSE .................................................................................................................. 10 Cognition - ADAS-cog .......................................................................................................................... 16 Function - ADCS-ADL ........................................................................................................................... 17 Behaviour - NPI ................................................................................................................................... 18 Global Status - CIBIC-plus .................................................................................................................... 19 Mortality ............................................................................................................................................. 20 Safety outcomes ..................................................................................................................................... 21 Serious Adverse Events (SAEs) ............................................................................................................ 21 Bradycardia ......................................................................................................................................... 22 Falls ..................................................................................................................................................... 22 Headaches ........................................................................................................................................... 23 Vomiting .............................................................................................................................................. 24 Diarrhea .............................................................................................................................................. 25

Ontario Drug Policy Research Network

5 Nausea ................................................................................................................................................ 26 DISCUSSION................................................................................................................................................. 30 ACKNOWLEDGEMENTS ............................................................................................................................... 33 REFERENCES ................................................................................................................................................ 34 Figure 1: Study flow .................................................................................................................................... 45 Table 1: Results of Network Meta-analysis for the Mini Mental State Examination (MMSE): all levels of severity of Alzheimer’s disease ................................................................................................................... 46 Table 2: Results of Network Meta-analysis for the Alzheimer’s Disease Assessment Scale ‒ cognition subscale (ADAS-cog): for all levels of severity of Alzheimer’s disease ....................................................... 50 Table 3: Results of Network Meta-analysis for the Alzheimer’s disease Cooperative Study activities of daily living inventory (ADCS-ADL) ............................................................................................................... 52 Table 4: Results of Network Meta-analysis for the Neuropsychiatric Inventory (NPI) ............................... 56 Table 5: Results of Network Meta-analysis for the Change plus Caregiver Input (CIBIC-plus)................... 58 Table 6: Results of Network Meta-analysis for Mortality ........................................................................... 60 Table 7: Results of Network Meta-analysis for Serious Adverse Events (SAEs).......................................... 62 Table 8: Results of Network Meta-analysis for Bradycardia ....................................................................... 65 Table 9: Results of Network Meta-analysis for Falls ................................................................................... 66 Table 10: Results of Network Meta-analysis for Headache ........................................................................ 69 Table 11: Results of Network Meta-analysis for Vomiting ......................................................................... 72 Table 12: Results of Network Meta-analysis for Nausea ............................................................................ 74 Table 13: Statistically Significant NMA Results for Mini Mental State Examination (MMSE): all levels of severity of Alzheimer’s disease ................................................................................................................... 77 Table 14: Statistically Significant NMA Results for Alzheimer’s Disease Assessment Scale ‒ cognition subscale (ADAS-cog): for all levels of severity of Alzheimer’s disease ....................................................... 78 Table 15: Statistically Significant NMA Results for the Alzheimer’s disease Cooperative Study activities of daily living inventory (ADCS-ADL) ............................................................................................................... 79 Table 16: Statistically Significant NMA Results for Neuropsychiatric Inventory (NPI) ............................... 80 Table 17: Statistically Significant NMA Results for Change plus Caregiver Input (CIBIC-plus) ................... 81 Table 18: Statistically Significant NMA Results for Mortality ..................................................................... 82 Table 19: Statistically Significant NMA Results for Serious Adverse Events (SAEs) .................................... 83 Table 20: Statistically Significant NMA Results for Bradycardia ................................................................. 84 Table 21: Statistically Significant NMA Results for Falls ............................................................................. 85

Ontario Drug Policy Research Network

6 Table 22: Statistically Significant NMA Results for Headache .................................................................... 86 Table 23: Statistically Significant NMA Results for Vomiting ...................................................................... 87 Table 24: Statistically Significant NMA Results for Nausea ........................................................................ 88 Appendix 1. List of Excluded Types of Dementia ........................................................................................ 89 Appendix 2. MEDLINE Search ..................................................................................................................... 90 Appendix 3. List of Included Studies ........................................................................................................... 95 Appendix 4. Study and Patient Characteristics ......................................................................................... 110 Appendix 5. Appraisal of bias of the included studies using the Cochrane risk-of-bias tool.................... 120 Appendix 6. Appraisal of bias of the included studies using the Cochrane Effective Practice and Organization of Care -of-bias tool (EPOC)................................................................................................. 125 Appendix 7: Appraisal of methodological quality of the included studies using the Newcastle-Ottawa Scale .......................................................................................................................................................... 128

Ontario Drug Policy Research Network

Rationale The efficacy of cognitive enhancer medications has been examined in previous systematic reviews [1-3]. These reviews have only included randomized clinical trials, (RCTs) which limit the ability to examine harms and generate ‘real world’ relevant findings. We are conducting this review to update a previous systematic review and network meta-analysis to determine the comparative effectiveness, and safety associated with cognitive enhancers versus, each other, no treatment, placebo, or best supportive care for patients with Alzheimer’s dementia. We will include data from all quantitative study designs, to provide more relevant findings for the ODPRN.

Methods Our systematic review protocol was drafted using guidance from the Preferred Reporting Items for Systematic reviews and Meta-analyses for Protocols (PRISMA-P) [4]. The protocol was circulated to various stakeholders, including policy makers from the Ontario Ministry of Health and Long-term care, industry partners, patients, researchers with the ODPRN, and health care professionals.

Eligibility criteria We included randomized clinical trials (RCTs) and non-randomized studies (e.g., controlled clinical trials, controlled before-and-after studies, cohort studies) examining the impact of cognitive enhancers (donepezil, galantamine, rivastigmine, and memantine) on cognition, function, behavior, global status, and mortality or harms amongst patients with Alzheimer’s dementia. Studies examining cognitive enhancers in any combination compared with each other, combinations of each other, no treatment, placebo, and best supportive care in patients with Alzheimer’s dementia were included. We excluded studies that examined any other cognitive enhancers or included mixed dementia patient populations. A full list of excluded types of dementia can be found in Appendix 1. Eligible studies had to report at least one of the following relevant outcomes: cognition (Mini– Mental State Examination [MMSE] [5]; Alzheimer’s Disease Assessment Scale ‒ cognition subscale [ADAS-cog]), function [5] (Alzheimer’s Disease Cooperative Study activities of daily living inventory [ADCS-ADL]), behavior (Neuropsychiatric Inventory [NPI]), global status (Clinician’s Interview-Based Impression of Change plus Caregiver Input [CIBIC-plus]), mortality, and harms (e.g., serious adverse events (SAEs), bradycardia, falls). According to previous studies, an increase of 3 points is considered a minimal clinically important difference (MCID) on the MMSE [5], while a reduction of 4 points on the ADAS-cog considered a MCID [6].These outcomes were selected by a panel of 32 clinicians who regularly care for patients with AD and policy-makers from Health Canada.

Information sources and literature search An experienced librarian updated our unpublished systematic review [7] by searching MEDLINE, EMBASE, the Cochrane Methodology Register, Cochrane Central Register of Controlled Trials, CINAHL,

Ontario Drug Policy Research Network

8 and Ageline databases from September 2011 until January 2015. The MEDLINE search is presented in Appendix 2. The results from the literature search were uploaded onto our online screening software (Sythesi.SR) [8].

Study selection process To ensure reliability, a training exercise was conducted prior to commencing screening. Using the inclusion and exclusion criteria, a random sample of 100 titles and abstracts from the literature search was screened by all team members during the level 1 screening (screening of titles and abstracts). This training exercise was then repeated for level 2 screening (screening of full-text articles). Subsequently, two reviewers screened citations for inclusion, independently, for level 1 and level 2 screening. Conflicts were resolved by discussion or the involvement of a third reviewer (HA, PR and ACT).

Data items and data abstraction process We abstracted data on study characteristics (e.g., country of conduct, study design, study period, study duration, setting), patient characteristics (e.g., AD diagnosis criteria and score, AD severity, age, sample size, comorbidities), and outcomes (e.g., sample size, mean, variance, number of patients experiencing harms). To classify the severity of AD for studies that did not report AD severity of the included patients, we used the following categories from the National Institute for Health and Care Excellence (NICE): MMSE of 21-24 as mild, MMSE of 10-20 as moderate, MMSE